[U-Boot] [PATCH] arm_cortexa8: support cache flush to other soc

Minkyu Kang promsoft at gmail.com
Tue Sep 8 02:06:04 CEST 2009


Dear Dirk,


>>>  I have two request.
>>
>> 1. I want to replace CONFIG_L2_OFF define to other function
>>
>> for example..
>>
>>    if (need_cache_flush()) { /* or !l2_off() */
>>        /* turn off L2 cache */
>>        l2_cache_disable();
>>        /* invalidate L2 cache also */
>>        v7_flush_dcache_all(get_device_type());
>>
>>        i = 0;
>>        /* mem barrier to sync up things */
>>        asm("mcr p15, 0, %0, c7, c10, 4": :"r"(i));
>>
>>        l2_cache_enable();
>>    }
>>
>>
>> 2. I don't want to use get_device_type() function
>> (just call like this: v7_flush_decahe_all() )
>>
>> How you think?
>>
>
> I wonder if these two requests are
>
> - nice to have topics? Or
>
> - required changes?
>


request no.1 is partly "required"
but not now, it is required for supporting s5pc110 soc.
because of we want to use same binary at runtime, so have to avoid ifdef
of cause, we can use other ways.. but i think it's good way that is removing
ifdef

request no.2 is just "request"
If you don't want to accept this, we can use the function.
but I think get_device_type() does not fit with cpu common file


>
> And if these are required changes, if
>
> - these can go on top of
>
> http://lists.denx.de/pipermail/u-boot/2009-August/058492.html
>
> so that we can apply above patch, you are able to go on to bring Samsung
> into mainline? And in parallel we discuss/change above topics? Or
>
> - we have to rewrite
>
> http://lists.denx.de/pipermail/u-boot/2009-August/058492.html
>
> what will stall Samsungs mainline integration further?
>
> Best regards
>
> Dirk
>
>
As a result, you can apply the patch.
but I hope we will process my request soon.

Thanks :)
Minkyu Kang

-- 
from. prom.
www.promsoft.net


More information about the U-Boot mailing list