[cap-talk] Midori in The Register

[Jonathan S. Shapiro]

On Wed, 2008-08-06 at 06:32 -0400, Kenneth Hamer-Hodges wrote:
> Plessey 250 system rejected ... any hardware DMA. 
> The multiprocessor architecture used any of the multiple CPUs to perform I/O
> actions directly by a thread with the requisite "capability". 

I think it's clear in hindsight that this was a bad call. If nothing
else, dedicating a CPU to data motion is a wasteful use of resource. An
alternative would be to implement a typed DMA mechanism at the hardware
interface. This would admit the possibility of a safe hardware DMA
subsystem, the remaining problem being the requirement to perform a safe
join following DMA completion.


shap