Re: Radeon KMS status (Re: [git pull] drm radeon kms.)

[Peter Zijlstra <peterz@infradead.org>]

On Wed, 2009-11-11 at 09:14 +0000, Dave Airlie wrote:
> > Hi Dave,
> > 
> > I was wondering about the current state of Radeon KMS, how far are we
> > from having it removed from staging?
> 
> I'm sort of thinking the next merge window of moving the enable option 
> into the kernel proper. We still have a metric shitload of work to get to 
> feature parity with current userspace drivers and we've done no 
> optimisation work on the rendering portion of the stack so I'm not sure 
> freezing the API isn't a bit premature, however maybe the API we have is 
> good enough to support and we can add new ioctls to make it go faster if 
> we require.
> 
> Jerome has proposed some API changes but we really haven't had time to 
> validate them in any useful way.

I by no means intend to rush things along, so if you think you need more
time, please take it ;-)

Just wanted to know where we stand... I'll try the Radeon KMS thing
again once I get through this mountain of email that accumulated over
the vacation and let you know how the current bits work.

Last time I tried KMS swapped the two DVI outputs wrt. the !KMS setup,
which is rather annoying during the transition period, but nothing I
can't live with once the stuff is stable enough to use all day.