[RS690] [X1200] graphic glitches (medium horizontal lines) when KMS on [KMS]

[Cutler]

Hello guys, I recently decided to give a try to new Ubuntu 10.04. Unfortunately I have a problem with my graphics as everything is messed up take a look on the screen attached. It might be something with the drivers, I have radeon X1200 and it is working ok only when using vesa driver. For the ati or radeonhd it's very bad. With 9.10 everything worked perfectly.

PS: I solved it by turning the KMS off with the nomodeset option in the boot options.

Screenshots: http://img402.imageshack.us/i/screenshot1dm.jpg/ , http://img402.imageshack.us/i/screenshotkt.jpg/

[lspci]
00:00.0 Host bridge [0600]: ATI Technologies Inc RS690 Host Bridge [1002:7910]
        Subsystem: Micro-Star International Co., Ltd. Device [1462:2fbd]
01:05.0 VGA compatible controller [0300]: ATI Technologies Inc RS690M [Radeon X1200 Series] [1002:791f]
        Subsystem: Micro-Star International Co., Ltd. Device [1462:2fbd]

[James P. Carter (jpcarter)]

I changed the package selection to xserver-xorg-driver-ati.

[Bryce Harrington (bryce)]

Hi radim-vajchr,

Please attach the output of `lspci -vvnn` and `dmesg`, and attach your /var/log/Xorg.0.log (and maybe Xorg.0.log.old) file from after reproducing this issue. If you're using a custom /etc/X11/xorg.conf please attach that as well.

[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[Cutler (radim-vajchr)]

00:00.0 Host bridge [0600]: ATI Technologies Inc RS690 Host Bridge [1002:7910]
        Subsystem: Micro-Star International Co., Ltd. Device [1462:2fbd]
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap- 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort+ >SERR- <PERR- INTx-
        Latency: 0

00:01.0 PCI bridge [0604]: ATI Technologies Inc RS690 PCI to PCI Bridge (Internal gfx) [1002:7912]
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B- DisINTx-
        Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 64
        Bus: primary=00, secondary=01, subordinate=01, sec-latency=64
        I/O behind bridge: 0000c000-0000cfff
        Memory behind bridge: fdd00000-fdefffff
        Prefetchable memory behind bridge: 00000000d0000000-00000000dfffffff
        Secondary status: 66MHz+ FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- <SERR- <PERR-
        BridgeCtl: Parity- SERR+ NoISA- VGA+ MAbort- >Reset- FastB2B-
                PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn-
        Capabilities: <access denied>
        Kernel modules: shpchp

00:04.0 PCI bridge [0604]: ATI Technologies Inc Device [1002:7914]
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B- DisINTx+
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 64 bytes
        Bus: primary=00, secondary=02, subordinate=02, sec-latency=0
        Memory behind bridge: fdf00000-fdffffff
        Secondary status: 66MHz- FastB2B- ParErr- DEVSEL=fast >TAbort- <T...

[Bryce Harrington (bryce)]

Hi Cutler,

From your screenshots, it definitely looks like a mode-setting error. With 10.04, mode-setting for -ati is being done in the kernel rather than in X.

Recently, the kernel team backported the 2.6.33 drm into Ubuntu Lucid. We're finding this resolves a huge number of problems. I don't know if it solves yours but bet it does. Would you mind updating to the latest Lucid and re-testing?

If it does still happen, please run the command:

  apport-collect 530725

This will attach all the logs and associated files we may need for troubleshooting this issue.

[Psy[H[] (vovik-wfa)]

On Lucid beta with radeon HD 2400 and default -ati driver I have minor glitches:
randomly appearing short (~20-30 pixels) horizontal white stripes. It looks like analog interferences, barely noticeable, but present.

[Bryce Harrington (bryce)]

[Resetting to incomplete]

[Rainer Oliveira (raineroliveira)]

Hi all.

I'm too seeing small corruptions among the screen.
I'm on a laptop with Radeon HD 4200.

It's just like Psy on comment #6 said. it looks like some kind of interference and it's barely noticiable, but it is there (it is noticiable specialy on plymouth boot screen, where the screen is lighter).
In Karmic and in Lucid Aplha 3, I had no problems. It seems it has started in Lucid Beta or maybe a little before (cant say the right date, though). So I think it could be a regression.

Thanks.

[Bryce Harrington (bryce)]

[Resetting to incomplete; we need comment from cutler]

[Brian Ealdwine (eode)]

I'm pretty sure that Psycho and Rainer have the same problem I do, and that it's the problem Cutler had.

First: I've upgraded from 9.10 to 10.04 beta. My system updates are current.
  I have a radeon rs690m (x1200)

This is the problem I ran into:
Small, horizontal white lines that appear a bit like static. Not very noticeable.
Medium and large horizontal lines / areas of corruption.
  * Stripes are often screen-wide
  * Stripes can be scrolled, e.g., in Firefox or Nautilus.
  * Refresh of graphical object clears up corruption -- E.g., a mouseover of a link or highlighting an icon
Mouse Pointer gets corrupted at some points, until refreshed (e.g., when the cursor changes from pointer to question mark, or to a hand, or from hand to arrow, etc.).
  * Corruption occurs regardless of whether or not Desktop Effects are enabled (important later)

I don't know if this is the same issue that cutter is having, and without feedback from Cutter, I won't. But, I would like to see this bug resolved before the release.

Resolution (ish):
In /etc/modprobe.d/radeon-kms.conf, I set:
 options radeon modeset=0

Why just "ish"?
  * Logging into an account created when the system was 9.10 fails, at best booting me back to the login screen.
  * Logging into an account created since the system has been upgraded to 10.04 works fine, desktop effects and all, no graphics glitches.
  * Logging into the 9.10-made account with desktop effects disabled works fine.
  * Enabling desktop effects once logged in makes the system visually unusable -- cannot switch to console, graphical screen unresponsive except for mouse.

Note that although there were glitches graphically, desktop effects or not, the system worked with desktop effects enabled when kms was enabled.

I think there are probably two problems here:
1) KMS causing or making visible some kind of horrible memory badness that I don't understand
2) Desktop Effects config in 9.10 can be set to a state that doesn't play nicely in 10.04 with KMS disabled.

Any further information I will be happy to provide. I have a workable system, but I want to make sure others aren't affected problematically.

-Brian

[Bryce Harrington (bryce)]

Brian, the easiest way for us to handle this is if you file a new bug. It's easier for us to handle two separate bug reports that end up being dupes than it is to handle one bug report with multiple potentially-unrelated issues in them.

Please file it using the command `ubuntu-bug xorg`, which will ensure necessary files are automatically attached (without those files I have no chance of determining if you are seeing the same bug or not). Also since this is a corruption bug, we'll want to have a photo of the screen (screen captures don't usually work, so you need to use a camera).

Meanwhile, I'll reset to incomplete to wait for cutter's reply on his bug.

[Brian Ealdwine (eode)]

Will do.

[Brian Ealdwine (eode)]

Posted bug #556782 for the issues I posted in this bug that may be a separate issue.

[Bryce Harrington (bryce)]

We're closing this bug since it is has been some time with no response from the original reporter. However, if the issue still exists please feel free to reopen with the requested information. Also, if you could, please test against the latest development version of Ubuntu, since this confirms the bug is one we may be able to pass upstream for help.

[Cutler (radim-vajchr)]

Sry guys for not aswering you for such a long time, but I was a bit busy. Anyway updating to the newer kernel did the trick and now everything works perfectly smooth and nice.