On Sat, 2004-09-25 at 07:19, Ronny V. Vindenes wrote:
> lør, 25,.09.2004 kl. 15.00 +0200, skrev Dieter Nützel:
> > > fre, 24,.09.2004 kl. 14.06 +0200, skrev Dieter Nützel:
> > >
> > > >
> > > > Latest DRI and DRM CVS.
> > > >
> > > > Yesterday's DRI CVS was fine (textures).
> > >
> > > I see this too, but not only in ut2k3/4 - I see it in several programs,
> > > but often it doesn't happen immediately e.g. neverball is fine untill I
> > > enter the option screen - then all textures are broken untill I restart
> > > the game.
>
> cvs as of today is almost fixed - the textures no longer disappear
> completely, but sometimes they flash off-on. I see it in both ut2k4 and
> neverball
So, it seems like we're seeing more ordering issues, now that things get
emitted less often. Moving the eye emit to the end mostly fixed
neverball's intro screen for me except for occasional flicker, and
in-game seemed OK. Patch attached. Do we have any systematic way to
figure this stuff out?
R100 seems OK with neverball.
--
Eric Anholt [EMAIL PROTECTED]
http://people.freebsd.org/~anholt/ [EMAIL PROTECTED]
Index: r200_cmdbuf.c
===================================================================
RCS file: /cvs/mesa/Mesa/src/mesa/drivers/dri/r200/r200_cmdbuf.c,v
retrieving revision 1.11
diff -u -r1.11 r200_cmdbuf.c
--- r200_cmdbuf.c 24 Sep 2004 04:20:58 -0000 1.11
+++ r200_cmdbuf.c 25 Sep 2004 22:10:22 -0000
@@ -94,7 +94,6 @@
insert_at_tail( &rmesa->hw.atomlist, &rmesa->hw.cube[i] );
for (i = 0; i < 3 + mtu; ++i)
insert_at_tail( &rmesa->hw.atomlist, &rmesa->hw.mat[i] );
- insert_at_tail( &rmesa->hw.atomlist, &rmesa->hw.eye );
insert_at_tail( &rmesa->hw.atomlist, &rmesa->hw.glt );
for (i = 0; i < 2; ++i)
insert_at_tail( &rmesa->hw.atomlist, &rmesa->hw.mtl[i] );
@@ -104,6 +103,7 @@
insert_at_tail( &rmesa->hw.atomlist, &rmesa->hw.ucp[i] );
for (i = 0; i < 6; ++i)
insert_at_tail( &rmesa->hw.atomlist, &rmesa->hw.pix[i] );
+ insert_at_tail( &rmesa->hw.atomlist, &rmesa->hw.eye );
}
void r200EmitState( r200ContextPtr rmesa )
