Re: [Dri-devel] radeon dri + software suspend/resume

[Charl P. B. <c.p...@it...>]

On Fri, Aug 02, 2002 at 05:27:46PM +0200, Michel Dänzer wrote:
> Okay, I've cleaned this up a little and extended it so that it actually
> survives if it can't reinit the DRI in EnterVT. :)
> 
> http://penguinppc.org/~daenzer/DRI/radeon-reinit.diff
> 
> The only possible issue I see is that the GLX visuals are only
> initialized on server startup. Could that cause apps to fail if the
> server has initialized the DRI on startup but it's not available when
> the app starts?
> 
> Otherwise, I think I could commit this right away, but it might be
> worthwhile to create a branch first to get some more testing and improve
> it further (I'd like to merge the 2D acceleration code, ..., other
> drivers?). What do you all think?

I've tested this on my Radeon M7 and on my setup it's rock-solid.  This is
in a different league than my first attempt (the driReInitKludge).

For those who don't know the details yet, here is an example:
1. I start two DRI X-servers, X1 and X2.
2. I can switch between these two without problems, running DRI clients on
   either one
3. If there's a DRI client running on X2 when I switch back to X1, X1
   automatically switches back to SW rendering.
4. I can now switch back to X2, stop the DRI client, and switch back to X1,
   which will now regain DRI.
5. As a a byproduct advantage (and this is what I did the kludge for in the
   first place), I can suspend to disc and resume back into a DRI X!
   
I will make binaries of the DRI snapshots with Michel's patch integrated
available at http://cpbotha.net/dri_reinit_kludge.html (just because this
page already exists and some people are using the kludge) so that some
feedback can be generated.  However, I really believe that this is NOT a
kludge anymore and that it should be integrated in CVS.

Thanks Michel!

Best regards,
Charl

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/