Re: Problem with FB

[Nicolas Ferre <nicolas.ferre@rfo.atmel.com>]

Diego A. Fons :
> Nicolas Ferre escribió:
> 
>> Diego A. Fons :
>>  
>>
>>> Hi,
>>>
>>> First of all i'm not sure if it is the correct list for posting this 
>>> mail, if it's not, please tell me wich list is the correct one.
>>>
>>> I'm working on an embedded device (a Ronetix pm9261 board with an 
>>> AT91SAM9261 cpu), i could build the linux kernel version 2.6.19.7 
>>> with framebuffer support and it works fine. The problem is when i run 
>>> any command that uses net services, i.e. when i run sftp the sreen 
>>> moves and it kept like this, the offset (0,0) is in position (100,0) 
>>> (it's just an example, i don't know thw correct values) and i'm not 
>>> able to correct it. It occurs every time i run a net comand. The 
>>> display i'm usin is a LCD and the driver is sidsab.c.
>>>   
>>
>> Strange behavior is not it ;-)
>>
>> I have seen this before and it was due to a bad timing setup of the
>> dm9000 chip select.
>>
>> -    at91_sys_write(AT91_SMC_CYCLE(2), AT91_SMC_NWECYCLE_(0) | 
>> AT91_SMC_NRDCYCLE_(16));
>> +    at91_sys_write(AT91_SMC_CYCLE(2), AT91_SMC_NWECYCLE_(16) | 
>> AT91_SMC_NRDCYCLE_(16));
>>
>> Is you have the first configuration, try with the second and it should
>> fix your problem.
>>
>> Hope that it helps. Tell us if it works.
>>
>> Regards,
>>  
>>
> Thanks for replying!
> 
> I search the line you describes and i could find something similar in 
> the initialization routine (attach) this is the similar line:
> 
> at91_sys_write(AT91_SMC_CYCLE(2), \
>                        (       AT91_DM9000_NWE_CYCLE |
>                                AT91_DM9000_NRD_CYCLE));
> 
> The macros are:
> 
> #define AT91_DM9000_NWE_CYCLE           (32 << 0)
> #define AT91_DM9000_NRD_CYCLE           (32 << 16)
> 
> Did you refer to that? or i have to look another place?

Definitely yes. (sorry I should have said that is was in board-xxxx9261.c
kind of files).

In this case AT91_DM9000_NWE_CYCLE is... 32 ! So my patch does not apply
(we had to set it to 16 instead of 0 on our sam9261ek eval board).

I think that ronetix is running the sam9261 at a higher speed than we do on
the Atmel sam9261ek : you should then reconsider the dm9000 timings (though they
seem to have adapted theirs already...).

Tell us if you find the proper ones (or if it is another reason).
Regards,
-- 
Nicolas Ferre



-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/