On Thu, Oct 25, 2001 at 08:10:09AM -0600, Keith Whitwell wrote:
> Bill Currie wrote:
> > i--;
> > list[0].prev = MGA_NR_TEX_REGIONS;
^^^^^^^^^^^^^^^^^^^
> > - list[i].prev = i-1;
> > + if (i)
> > + list[i].prev = i-1;
> > list[i].next = MGA_NR_TEX_REGIONS;
> > list[MGA_NR_TEX_REGIONS].prev = i;
> > list[MGA_NR_TEX_REGIONS].next = 0;
....
> Hmm. That list should be circular, so this patch isn't correct, but there may
> be a bug there for some other reason. I'll look at it...
That's just it, prior to my patch, that circularity was being broken. Whenever
i was zero, the previous list[0].prev = MGA_NR_TEX_REGIONS; was getting
torched with i-1 -> -1 -> 255. With my patch, if i==0, the list[0].prev is
left with the MGA_NR_TEX_REGIONS which completes the circle. I suspect the
only reason I'm getting i==0 is my measly 8M card running at 1152x864x16.
That said, I understand, accept and agree with your need to look at it more
closely :)
Bill
--
Leave others their otherness. -- Aratak
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel
