On Die, 2003-04-22 at 23:48, Leif Delgass wrote:
> On Tue, 22 Apr 2003, Leif Delgass wrote:
> > On 22 Apr 2003, Michel Dänzer wrote:
> > > On Die, 2003-04-22 at 22:53, Leif Delgass wrote:
> > > > Shouldn't PCIGART_ENABLED be removed from radeon_cp.c now, so pcigart
> > > > support will be included for all arches?
> > >
> > > Yes, it was removed in XFree86 CVS before 4.3.0 so it's a merge error if
> > > it's still there.
> > >
> > > > Also, the dependency on AGP for the radeon module in Config.in and Kconfig
> > > > should probably be removed.
> > >
> > > Absolutely.
> > Ok, I've fixed both of these in the DRI trunk.
> > It appears that the PCIGART_ENABLED define is still in the drm in XFree86
> > CVS as well, so it wasn't a merge error. The define was removed from
> > the DDX in both XFree86 and DRI CVS however.
> Actually, it looks like it _was_ a merge error. PCIGART_ENABLED was
> removed from the XFree86 CVS radeon drm, but was added back in the sync
> David did a few days ago.
I suspected so. BTW, I can recommend meld for anyone doing a merge, it's
pretty nice, a three way radeon_driver.c merge pretty much killed it
performance wise though. :)
> I noticed another problem with that merge. Charl Botha's DRI resume
> patch was committed to XFree86 CVS, which adds an ioctl to the radeon
> DRM. The DRM portion of that patch appears to have been reversed in
> the sync. I wonder if we should apply that patch to the DRI tree to
> get synced up with the ioctls?
Yes, and bump the minor version, and preferrably refactor the code a
little to reduce duplication, ...
Earthling Michel Dänzer \ Debian (powerpc), XFree86 and DRI developer
Software libre enthusiast \ http://svcs.affero.net/rm.php?r=daenzer