On Wednesday 16 March 2005 05:25, Timothy Miller wrote:
> On Wed, 16 Mar 2005 00:43:37 +0100, Nicolai Haehnle <[EMAIL PROTECTED]> 
wrote:
> > I agree. I don't know Timothy's precise reasoning, but I could
> > imagine that using PIO for the cursor is better for latency.
>
> No.  It's just one less thing that the DMA engine doesn't have to be
> able to talk to.  Mind you, that depends on the design.  I'm just not
> making any promises about optimizing such a trivial thing.
>
> Even if you CAN do cursor by DMA, why bother?  It won't save much
> time, because it's not something that's done often or in bulk.  Plus,
> you may want to change it on a vertical interrupt (to avoid possible
> video glitches), which means you're in a different context from the
> one that manages the ring buffer, and you don't want to have to mess
> with locks or deal with race conditions.  Doing it by PIO is entirely
> modeless.

Yah, that was me being stupid.  You're entirely right about that.

Regards,

Daniel
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to