[Justification]
During Intrepid development, we worked with upstream to develop two patches to work around banding issues on certain classes of hardware. Upstream has carried on this development work and corrected several flaws that were subsequently found in our patches (including regression bug #274234).
[Development]
Upstream's git tree contains the fixes for this. We will be pulling them in when we next update our git snapshot and will be included in the next -ati release.
[Patch]
debdiff of changes cherrypicked from upstream git is attached below.
In the development of our original fixes, we had found it difficult to insert our fixes due to the way the code was organized. To correct this, upstream restructured the code that handles dithering. This is why the patch is quite a bit larger than our original patches.
[Test Case]
On some HD 2600 hardware, booting up into X results in an X failure with the following error message in Xorg.0.log:
(EE) RADEON(0): ParseTable said: CD_INVALID_OPCODE
[Regression Potential]
The size of this patch alone suggests that regression potential is a concern. A code review did not turn up any obvious flaws, and preliminary testing by myself (and others who have tested snapshots of the upstream code) shows it to be safe. However, I think this should be kept in -proposed for at least a couple weeks. I'd like to see it tested across a wider range of ATI hardware.
[Original Report]
Currently the xserver-xorg-video-ati 1:6.9.0+git20081003.f9826a56-0ubuntu2 is shipping two patches to workaround dithering problem on some cards. I suggest to update to current git version which has proper fixes for this problems:
* RV530 on MacBook Pro
+ bug #264535
+ https://bugs.freedesktop.org/show_bug.cgi?id=17897
+ the current upstream fix has the following advantages over current ubuntu 103_rv530_enable_diethering.patch:
- it fixes completely the problem (the ubuntu patch still leaves some banding);
- it is targetting only RV530 on MacBook Pro (where there is the dithering problem), not all RV530;
- it makes the 103_rv530_enable_diethering.patch useless making the ubuntu packages more inline with supported upstream;
+ I tested current git version and noticed no regression so far
* RV6xx/RV7xx cards:
+ https://bugs.freedesktop.org/show_bug.cgi?id=17094
+ current git version should also fix this problem in a similar way, though I don't own this hardware to test; it should have the following advantages:
- it should fix completely the problem (the ubuntu patch appears to leave some banding);
- it has a more general approach targetting not only specified cards;
- it makes the 101_rv635_enable_dithering.patch useless making the ubuntu packages more inline with supported upstream;
Current git version (2008-10-15 commit 435cf7da68186f2601c4b888296117d4f652c625) has 9 commits more than current ubuntu version:
- 8 commits related to the dithering problems;
- 1 cleanup commit renaming some function;
I suggest to update the ubuntu version to current git to properly fix the dithering problems.
Test packages provided by Tormod Volden can be found at:
https://edge.launchpad.net/~tormodvolden/+archive
Hi fabio-pedretti,
Please attach the output of `lspci -vvnn`, and attach your /var/log/Xorg.0.log file from after reproducing this issue. If you've made any customizations to your /etc/X11/xorg.conf please attach that as well.