El Dom 14 Oct 2001 14:32, Jouko Pynnönen escribió: > I got an Acer Travelmate laptop this week and would like to see 3d > acceleration on it, so i tried tweaking with the mach64 branch. > > There was a minor issue with up() and down() which caused a kernel oops. > After fixing that, I commented out some code in mach64_do_dma_init() which > doesn't seem to work. The clear code was disabled with an #if 0. After > enabling that, i can run e.g. gears and it looks ok although the fps rate > is relatively low (below 300). You're right. The problem is that the clear code was into a #if0.- #endif After enabling the code, I've also got 210-220 fps with the 4.1.0 >
> > What's next? What is the "BusMastering trouble"? Are there any specs > available for wannabe dri developers, or are we to guess what mach64 > registers do? Any other recommended docs to start with? The code you've commented out in the mach64_do_dma_init is a test for making bus mastering. The code you're testing is really making direct register programming, what is not a good thing. The target is to make the register writes using DMA Busmaster transfers (you can look at the original DMAOUTREG, DMAADVANCE macros in the Utah-GLX code), but the test for the Busmaster transfer is failing and locking the card engine all the time. You can register in ATI developers program. It seems that lately they're releasing the register docs to people asking for them. BTW, I'm trying to adapt the code to the DRI trunk to work with a more modern codebase. Unfortunately, the system is still hanging. -- M. Teira _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
