Eric Anholt wrote: > On Thu, 2002-07-04 at 00:10, Keith Whitwell wrote: > >>Eric Anholt wrote: >> >>>Here's my current list of concerns about the state of both the branch >>>and the general CVS support of Radeons (since that's what another user >>>and I are both playing with the most). >>> >>>BSD branch's fault >>>- I haven't cleaned up the non-ostemplated <driver>_drv.c files in >>>linux, which means they will have a multiply defined symbol when >>>compiled into the linux kernel. >>> >>Can you detail this a bit more? >> > Jens was talking about cleaning up the <driver>_drv.c files, so I did it > part of the way for the os-templated ones. If you look on bsd-branch, > they are just header includes now on Linux. Before there were several > other things, including the module init code that was put into each file > because of preprocessor troubles (macros as macro args). I did (I'm > pretty sure) the necessary preprocessor magic and put it in drm_drv.h in > linux, trivially tested by compiling the code. However, I didn't > compile the other drivers, and forgot about it because the other drivers > (i810) aren't compiling for me on linux. They'll have it defined in > drm_drv.h and in <driver>_drv.c > > >>>- MGA is still slower on BSD than Linux (probably a historical problem, >>>so nothing to block a merge). >>> >>>probably not BSD branch's fault: >>>- When two glxgears are run and dragged to be overlapping, there are >>>obvious clipping issues. >>> >>In pageflipping mode? Or generally? >> > Generally. It was 100% of the time on the testbox with twm running. > The bottom gears showed up in the top one. This was non-pageflip. > Another user noticed this with pageflip enabled. > > On my main machine (enlightenment, opaque dragging, 7500 rather than > VIVO) the bottom gears would flicker in and out of the top gears's > window. Also non-pageflip. > > Also, I just remembered, sometimes on the testbox dragging the top gears > didn't move where the top gears were located, just showed more black and > the two sets of gears (top + bottom) in the same places they were > pre-drag. Haven't looked into this much, not sure if it even required > two gears running, and didn't happen all the time anyway. > > Make any sense?
These all seem to relate to the SET_SCISSORS patch that went in recently... It seems that more sychronization is required or the scissors uploaded that way are ignored by the 3d part... Tim? Are you seeing these droppings? Keith ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Caffeinated soap. No kidding. http://thinkgeek.com/sf _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
