Xrender cursor performance / double buffering?

[Jim Gettys]

As Dan was saying, we'll revisit the whole situation with the LX
hardware.

Even if we keep most applications using 16 bit depth (to save memory,
bandwidth and power), we can possibly move to a 8888 ARGB root window,
and composite the applications' data (and cursor).

But I'd like some real data first ;-).  This isn't an obvious slam dunk
decision to make, and I'd like it based on data we don't have yet.

                               - Jim


On Tue, 2007-04-17 at 12:55 -0400, Adam Jackson wrote:
> On Tue, 2007-04-17 at 13:07 -0400, Jim Gettys wrote:
> > Jordan,
> > 
> > One can compute the cursor image once to match the depth of the root
> > window once and store it in devprivates.  Then the server can composite
> > the image using the fast path image compositing hardware, no matter what
> > the depth is (it won't matter if computing the cursor to match the
> > screen takes a bit of time; it's the compositing while the cursor moves
> > that really hurts).
> 
> That won't work if the root window depth is 16.  You won't have any
> place to store your alpha channel.  You can do a transparency blit, but
> then your smooth edges go away.
> 
> Also, the software alpha cursor code in X sucks out loud.  It'll use the
> normal Render path, yes, but it still blinks violently.
> 
> - ajax
> 
-- 
Jim Gettys
One Laptop Per Child