From: Andreas Beck <becka@rz.uni-duesseldorf.de>
  To  : ggi-develop@eskimo.com
  Date: Tue, 17 Aug 1999 00:07:23 +0200

Re: Ping-pong buffers on KGIcon are here!

> Why not require the application (and GGI-lib functions)
> to make a check and sleep if accels may be running still, as
> suggested earlier in this thread?

That should be the case. See my other mail on advisory locking.

> This may require a change in the GGI-API, 

It doesn't. the API is there. Direct fb access has functions to
lock/unlock it. And the LibGGI rendering libs should use them, too.

> this point, but personally, I don't like the idea of
> unmapping/mapping the framebuffer 

Yes. This should only be done, if it can lock the system. IMHO on these
cards, I'd rather keep to synchronous accels, avoising the problem, and
probably still having better performance.

> because of the fact that a program can't know in advance if it will 
> run into performance trouble or not - this is dependant on the
> underlying video-hardware and driver, and how well GGI
> makes use of that driver.

Exactly. Block/Line combinations will be perfectly o.k. for one card, and
poisonous for another one (that can't do line in HW).

CU, ANdy

-- 
= Andreas Beck                    |  Email :  <andreas.beck@ggi-project.org> =