>    Does the change break on other Apple laptops? 
   > do you not know that already?  ie, i'd hope you had wider testing of
   > both new & old (ie, already working) hardware before commiting tis.
   
   This is a question...

... that you should answer :-)
   
   > yeah i'd like it to be an XF86Config option that you have to set, or
   > at least derived someway and only enabled where it is known needed,
   > or something.
   > 
   > i don't particularly understand the problem itself - i'm objecting
   > to it being applied with a hammer to all powerpc platforms when it's
   > (a) a complete hack (please clean up the magic number at least) and
   > (b) it's already working in older systems...
   
   Then what about the second solution I proposed: An ibookradeon driver.
   ibookradeon_driver.c will just do that:
   
   #define IBOOK
   #include "radeon_driver.c"
   
   And we'll ifdef IBOOK the changes in radeon_driver.c

i don't think this is supportable.  there's far more than
radeon_driver.c in the radeon driver itself...



.mrg.