Re: [PATCH] sparc32,leon: need VIRQ:IRQ 1:1, don't mask/ack IRQ Controller

[Daniel Hellstrom <daniel@gaisler.com>]

David Miller wrote:

>From: Daniel Hellstrom <daniel@gaisler.com>
>Date: Wed, 30 Mar 2011 14:41:45 +0200
>
>  
>
>>I see, however that is almost as bad because that would also require
>>one driver per core but in the LEON IRQ layer of Linux.
>>    
>>
>
>This is absolutely unavoidable since the firmware device tree doesn't
>pass along the proper set of IRQs.  You have to fixup your tree,
>because it is broken.
>  
>
This is sad, as we have designed all LEON architecture to be Plug&Play 
as far as possible, and it will be software that limits it in the end. I 
think it is quite uncommon (if not unique) to do plug&play on the most 
low-level stuff such as timers and IRQ Controller etc., I believe that 
part of the idea with the OpenBoot tree is to provide that low-level 
plug&play to Linux and LEON that has true PnP in hardware can not do 
that fully. I'm starting to think virtual IRQ is a bad thing.. :)

I will give it another thought tomorrow before doing anything. A part of 
me thinks that the bootloader should be modified to report the correct 
information, however that requires changes to u-boot as well.

Daniel