Patch for bug #1912 applied incorrectly to 6.8 branch

[Michel Dänzer]

On Wed, 2005-02-09 at 22:01 -0800, Alan Coopersmith wrote:
> 
> It looks like the two patches (bug #1912/patch #1433 and bug 
> #1220/patch #980) somehow got merged into this single commit, and that the 
> changes to radeon_driver.c between the time the diff was made for patch #980
> and when the patch was applied (including the original commit of the fix #980
> back in September) caused the patch #980 portion of the change to appear in
> the wrong section of the code the second time around.

The hunk context doesn't even look similar to me though...

> Of course, that's mostly academic, and the real question is how bad is the bug
> introduced by the extra line of code?   Bad for performance?   Likely to cause
> crashes or hangs?

As I said before: it's unnecessary writes to GPU memory controller
registers. In the best case, there's an unnecessary slight delay maybe;
in the worst case, the GPU might wedge.


-- 
Earthling Michel Dänzer      |     Debian (powerpc), X and DRI developer
Libre software enthusiast    |   http://svcs.affero.net/rm.php?r=daenzer