RandR 1.3 additions?

[Alex Deucher]

On 7/17/07, Matthias Hopf <mhopf at suse.de> wrote:
> On Jul 16, 07 12:43:42 -0400, Alex Deucher wrote:
> > > Are there other things we should consider adding to the RandR protocol
> > > at this point?
> >
> > How about GPU and CRTC properties?  It would be nice to have a
> > standard interface to adjust global GPU options (e.g., clock gating
> > and other power management stuff, GPU and memory clocks, framebuffer
> > compression, tiling, etc.).  The CRTC properties could be useful for
> > the projective transforms and on a lot of chips the panel scaler is
> > actually part of the crtc rather than the output.
>
> While I very much agree that we need a mechanism for things like that
> (include antialiasing, anisotropic filtering, vblank, etc. here), I'm
> not entirely sure that this fits the RandR extension. Many of these
> parameters are not output dependent, but card and/or host dependent
> (read the later: global).

I see randr as more of a dynamic configuration control than
specifically for outputs (although that's what it focuses on now).  I
think it would sense to add this to randr than to add a new extention
for this sort of thing.  We'll also need some concept of a GPU for
supporting multi-card with randr and xinerama anyway.  We already have
dri config and xv attributes and now xrandr; any true GPU config tool
would need to deal with all of them.  They are all basically just
attribute systems with driver callbacks.  Why add another?

Alex