quoted
There's been a patch by Geert on this list to fix this - basically
subtract GUI_RESERVE from the determined vram size in atyfb.c for all 8 MB
Rage Pro (and possibly other) chips to prevent the screen clear from
stepping on the card registers.
It's a different issue. This machine has a r128 and aty128fb works fine
with the few fixes we added for the M3. The problem is with offb and seem
to be related to some timing issues when accessing the DAC palette
registers. Henry found a fix for this, I'll add this to my kernel asap.
OK, I didn't catch the point that the new PB has a rage128. You're lucky,
then.
Michael
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/