From: Leigh D. <lsd...@ec...> - 2003-07-25 03:53:53
|
On Fri, 2003-07-25 at 02:25, Michel Dänzer wrote: > On Thu, 2003-07-24 at 04:50, Leigh Dyer wrote: > > > > The radeon-accel-init.diff patch seemed to work okay at first - the X > > server started normally, but with DRI disabled, as expected. I've > > attached my X log confirming this. > > > > After applying the drm-agp-acquire.diff patch, the X server begins to > > load, and the log shows that AGP initialisation fails, the system drops > > back to PCI mode, and DRI is enabled. However, initialisation stops at > > the line: > > > > (II) RADEON(0): Acceleration enabled > > > > I've attached the log of this session as well. I also happened to > > accidentally test the server using a DRM module compiled from your > > packaged sources, and it stopped in the same place (though I didn't > > think to check whether or not AGP or DRI were successfully enabled). > > Weird. It would be interesting to see the log from that, as well as what > happens with only drm-agp-acquire.diff applied but without > radeon-accel-init.diff . Thank you for your testing. > I just tried things with the drm-agp-acquire patch but without the radeon-accel-init patch - X came up without a problem, and DRI works, albeit slowly. The log confirms AGP initialisation fails cleanly, and DRI is successfully enabled in PCI mode. Using a DRM module compiled from your packaged sources along with my compiled DRI CVS seems to work very well. The DRM module reports itself as version 1.8.0, while the DRM module in DRI CVS is 1.9.0 - is this worth worrying about for the moment? Thanks Leigh |