Bug#583604: xserver-xorg-video-radeon: Corrupt screen after playing Cytadela OpenGL game and then running Emacs
Hi,
On Sat, May 29, 2010 at 08:56:38AM +0200, Brice Goglin wrote:
> Le 29/05/2010 08:40, Sylvain Beucler a écrit :
> > reopen 583604
> > thanks
> >
> > To get the 2D performances back (300FPS vs. 17FPS) I had to disable
> > KMS. In this case, the display corruption still happens with the
> > newer driver and kernel.
>
> We're not going to track bugs for 6.13 without KMS, this case isn't
> recommended already and it won't be supported in the future.
Hmmm, you mentioned that 6.13 is what is planned to go in Squeeze.
As the maintainer of the freedink package, which is a 2D game, I
sincerely hope we'll find a way to provide good 2D performances by
default in Squeeze, either by fixing KMS performances, either by
fixing non-KMS display corruption. Otherwise this is a regression
against Lenny. Let me know if I'm mistaken.
I think this may be a Debian bug more than a Radeon bug.
What do you think?
I'm not rereopening to avoid an open/close war and wait for your
input.
> If KMS is slow, please open a different bug.
Sure, I'll do that too.
> Make sure you have firmware-linux installed (see in dmesg whether
> some firmwares are missing).
As I mentioned I installed the (sadly non-free) R300 2kB firmware.
Apparently dmesg is happy:
$ dmesg|grep -C2 firmware
[ 109.840767] [drm] Setting GART location based on new memory map
[ 109.840862] [drm] Loading R300 Microcode
[ 109.841585] platform radeon_cp.0: firmware: requesting radeon/R300_cp.bin
[ 109.909568] [drm] Num pipes: 1
[ 109.909581] [drm] writeback test succeeded in 1 usecs
--
[ 427.525752] [drm] Setting GART location based on new memory map
[ 427.525848] [drm] Loading R300 Microcode
[ 427.526558] platform radeon_cp.0: firmware: requesting radeon/R300_cp.bin
[ 427.536623] [drm] Num pipes: 1
[ 427.536636] [drm] writeback test succeeded in 1 usecs
Cheers!
--
Sylvain
Reply to: