Created attachment 48940 [details]
xorg.conf

Created attachment 48941 [details]
Xorg.0.log

Created attachment 48942 [details] [review]
dmesg

What kernel are you using?  Is there some reason you are using UMS rather than KMS?  I'd recommend using KMS.

(In reply to comment #0)
> I tried to enable the new modesetting radeon
> driver in the 2.6.38-hardened-r6 kernel to see if that would help but it
> doesn't work with my HD4250 card yet.

It should work fine.  What sort of problems are you having?

Created attachment 48947 [details]
"Welcome" text

Wow, KMS actually works fine.  It causes a delay of about 30 seconds or so during the kernel load which I had previously interpreted as a crash.

It doesn't fix the problem though.  I've attached another photo of an example of the problem.  The horizontal lines are photo artifacts but the "shadows" of the text to the left and right are on the screen.

I fixed that 30 second delay by loading the firmware properly.

Please attach your xorg log and dmesg output from KMS.  Does booting with radeon.audio=0 on the kernel command line in grub help?

Created attachment 48982 [details]
Xorg.0.log

Created attachment 48983 [details]
dmesg

Created attachment 48984 [details]
sample.jpg

radeon.audio=0 helps a lot.  I'd say the problem is reduced by about 75%.  I've attached the KMS dmesg and Xorg.0.log as well as a photo of the screen with radeon.audio=0.  Can it be cleaned up further?

Can you try the latest 3.0 kernels?

Would git-sources-3.0_rc6-r6 be OK?

yes.

I tried the 3.0 kernel with the same result.  Do you know what's causing the problem since the radeon.audio=0 setting improved it so drastically?

(In reply to comment #17)
> I tried the 3.0 kernel with the same result.  Do you know what's causing the
> problem since the radeon.audio=0 setting improved it so drastically?

radeon.audio=0 disables the hdmi data packets and runs the hdmi port in dvi mode.  I suspect your monitor doesn't like something about the hdmi packets.

In case it makes a difference, the monitor is a 47" LG HDTV.  Can we do anything else to reduce the amount of data sent up the cable, or maybe DVI mode is the bare minimum?  I tried 16-bit mode with no change.

(In reply to comment #19)
> In case it makes a difference, the monitor is a 47" LG HDTV.  Can we do
> anything else to reduce the amount of data sent up the cable, or maybe DVI mode
> is the bare minimum?  I tried 16-bit mode with no change.

DVI mode doesn't send any hdmi packets.  You might try playing with the settings on your TV or try disabling coherent mode:

xrandr --output HDMI-0 --set coherent 0

Tried both of those with no effect.  Do you think there could be some sort of electrical interference issue at play that is exacerbated by the HDMI packets?  I isolate the USB connection between my computer and USB sound card with an ADUM4160 chip and it improves the sound greatly because so much electrical noise travels up the USB cable to the DAC otherwise.

I suppose it's possible.

Anything else I could try?

Is it possible that the artifacts are from some kind of image filter (e.g. (un)sharpening, noise reduction) applied by the TV? Maybe it uses the presence / lack of HDMI packets and/or the mode timings as indicators for whether the source is a PC or video player...

I fiddled with every setting on the TV with no change.  Please let me know if you have any other ideas.  I'm happy to try different things.

Mass closure: This bug has been untouched for more than six years, and is not
obviously still valid. Please reopen this bug or file a new report if you continue to experience issues with current releases.