Re: [Dri-devel] gears problem

[Frank E. <fe...@ai...>]

On Thursday 04 October 2001 07:06, Michel Dänzer wrote:
> On Thu, 2001-10-04 at 03:56, Steven P. Lilly wrote:

> Everything else seems to work fine, which points at a glxgears bug?

I encountered something very similar with glxgears when I was attempting to 
get RagePRO DMA operation working for Utah-GLX under PPC.  

Manually sending the registers from the DMA buffer (psuedo-DMA) would work 
completely fine for glxgears, but bus-mastering mode would hang the box after 
about 5-10 frames.  More interesting was the fact that if I prevented the 
blue gear from being rotated (simple comment out of one line...) the 
bus-mastering mode worked fine.  At the time, I hadn't the time to pursue it, 
but it's been lingering in my mind since then because as soon as we get it 
working for x86, we're going to want to make it happen on PPC.  

I know, it's not the same chip, but someone could check and see if it's the 
same problem.

-- 
Frank Earl