[U-Boot-Users] ramdisk booting problem

[U-Boot-Users] ramdisk booting problem

[bott at dei.unipd.it]

Hi,

I am trying to boot linux with a ramdisk on a board based on TI dm6446.

I am facing a weird problem: it correctly boots images smaller than  
2.5MB, while bugger ones stop boot after image decompression, before  
the kernel boots.

I'm not able to find out the reason why it hangs.

Do you have any suggestion?

Thanks

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

[U-Boot-Users] ramdisk booting problem

[Markus Klotzbücher]

bott at dei.unipd.it writes:

> I am trying to boot linux with a ramdisk on a board based on TI dm6446.
>
> I am facing a weird problem: it correctly boots images smaller than  
> 2.5MB, while bugger ones stop boot after image decompression, before  
> the kernel boots.
>
> I'm not able to find out the reason why it hangs.
>
> Do you have any suggestion?

One common pitfall is that during decompression some vital code is
overwritten by the decompressed data. Double check that the memory area
to which you are decompressing your image is large enough to hold the
uncompressed data and doesn't overlap with any code.

Best regards

Markus Klotzbuecher

--
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-Users] ramdisk booting problem

[bott at dei.unipd.it]

Quoting Markus Klotzb?cher <mk@denx.de>:

> bott at dei.unipd.it writes:
>
>> I am trying to boot linux with a ramdisk on a board based on TI dm6446.
>>
>> I am facing a weird problem: it correctly boots images smaller than
>> 2.5MB, while bugger ones stop boot after image decompression, before
>> the kernel boots.
>>
>> I'm not able to find out the reason why it hangs.
>>
>> Do you have any suggestion?
>
> One common pitfall is that during decompression some vital code is
> overwritten by the decompressed data. Double check that the memory area
> to which you are decompressing your image is large enough to hold the
> uncompressed data and doesn't overlap with any code.

Thanks for you answer Markus,

this is what I suspect, but I don't have any idea how I can check it.  
I mean, how can I see where the decompressed image is put? I tried  
giving a look at lib/inflate.c to check how gunzip() works, but I'm  
not able to figure out where inflated data goes....

Bye,

Ottavio.


----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

[U-Boot-Users] ramdisk booting problem

[Markus Klotzbücher]

Dear Ottavio,

bott at dei.unipd.it writes:

> this is what I suspect, but I don't have any idea how I can check it.
> I mean, how can I see where the decompressed image is put? I tried
> giving a look at lib/inflate.c to check how gunzip() works, but I'm
> not able to figure out where inflated data goes....

Look at the code in common/cmd_bootm.c . IIRC the ramdisk is copied to
the end of RAM. The boot output should also give some hints. You'll need
to provide more information such as your bootlog, the commands you're
using to boot and your memory/flash setup for us to help.

Best regards

Markus Klotzb?cher

--
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-Users] ramdisk booting problem

[bott at dei.unipd.it]

Quoting Markus Klotzb?cher <mk@denx.de>:

> Dear Ottavio,
>
> bott at dei.unipd.it writes:
>
>> this is what I suspect, but I don't have any idea how I can check it.
>> I mean, how can I see where the decompressed image is put? I tried
>> giving a look at lib/inflate.c to check how gunzip() works, but I'm
>> not able to figure out where inflated data goes....
>
> Look at the code in common/cmd_bootm.c . IIRC the ramdisk is copied to
> the end of RAM. The boot output should also give some hints. You'll need
> to provide more information such as your bootlog, the commands you're
> using to boot and your memory/flash setup for us to help.

eventually yesterday I've been able to boot with initrd and not  
initramfs by leaving the kernel in the flash memory and coping the  
initrd in ram. With the exception of switching to initrd, the kernel  
is the same, so I really think it was a problem of overwritten memory.

I'll have to find out why booting the kernel from flash works and from  
ram not.

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.