RandR 1.2 feedback

[Keith Packard]

On Wed, 2006-11-29 at 15:17 -0800, Andy Ritger wrote:

> Yeah, I'll be interested to see what that code looks like.  I think to
> ensure the most flexibility, I'll want the NVIDIA X driver to be able
> to hook the RandR 1.2 requests at a very high level, but I'll look into
> the code and see.

Yeah, this code hooks right to the extension itself, so if it doesn't
work, the driver can always just not use it. The goal is to eliminate
all RandR knowledge from the driver so I can build the driver with old X
servers and still share the same startup configuration code. It's all a
complicated path to sharing the same code in the drivers and getting
both RandR 1.2 and startup configuration supporting all the same
options.

> Agreed that heterogeneous GPU environments are more challenging.
> It still seems like an OpenGL implementation problem, though.

I'd say it's a combination and will require exposing which part of the
screen is managed by which GPU. But, I get to solve it to support
multi-card GL...

> I'm still a little bit nervous about some of the corner cases, but I
> think for the common case it's workable. 

When you find a case that doesn't work, we'll fix it. I like that
approach better and better these days.

>  I'll try to dig into the code
> soon.  The master branch is the right place to track the changes, now?
> (I'm still a git noob)

I'm actually working on the randr-1.2-for-server-1.2 branch this week
while I wait for the input-hotplug code to stabilize enough so I don't
go mad every time I update.


-- 
keith.packard at intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.x.org/archives/xorg/attachments/20061129/f75960b2/attachment.pgp>