[PATCH] [POWERPC] Add docs for Freescale DMA & DMA channel device tree nodes

Kumar Gala galak at kernel.crashing.org
Wed Jan 23 09:00:16 EST 2008


On Jan 22, 2008, at 3:48 PM, Scott Wood wrote:

> Timur Tabi wrote:
>> On 83xx, all DMA channels share the same interrupt?
>
> Yes.
>
>> Couldn't we just specify the same IRQ in each channel's node, so  
>> that they look the same across 83xx, 85xx, and 86xx?
>
> The consensus that I thought we had reached was to let 83xx specify  
> the interrupt in the controller node (to make things simple for a  
> driver using the shared summary register rather than registering  
> four handlers for the same IRQ), but require that it also be in the  
> channel node, and that if there is an interrupt in the controller  
> node, that it be the same as in all the channels.

This was my understanding.. however we might not be documentation that  
clearly at this point.

>> My sound driver doesn't use Extended Mode, but it does check for  
>> "fsl,8610-dma-channel".  I'm thinking that maybe I should change to  
>> to look for "elo" or "eloplus", but it would be nice if we could  
>> make the DMA node for an 86xx SoC compatible with a driver that  
>> expects "fsl,elo-dma-channel".
>
> Sure it'd be nice, but it's not possible without going back in time  
> and convincing the hardware people to make them 100% compatible.

exactly.  You forget minor things like 83xx reg set is little endian  
and 85xx/86xx is big endian.

- k



More information about the Linuxppc-dev mailing list