Xv displays flickering pixel garbage (intrepid)

Binary package hint: xserver-xorg-video-ati

Since the xserver-xorg-video-radeon update from 1:6.9.0+git20080802.1f3eee36-1ubuntu1 to 1:6.9.0+git20080826.a3cc1d7a-2ubuntu1, all video players using Xv began displaying flickering pixel garbage. The corruption fills the overlay horizontally and fills a varying vertical area beginning from the bottom of the overlay.

The flickering seems to happen each time the frame is updated.

01:00.0 VGA compatible controller: ATI Technologies Inc RV350 AP [Radeon 9600]

I have the same problem with kubuntu intrepid alpha6, xserver-xorg-video-radeon version 1:6.9.0+git20080826.a3cc1d7a-2ubuntu3

Can you also attach the output of running lspci from the commandline? Thanks.

01:00.0 VGA compatible controller: ATI Technologies Inc RV350 AR [Radeon 9600]
01:00.1 Display controller: ATI Technologies Inc RV350 AR [Radeon 9600] (Secondary)

confirmation from here... exact same issue, xv unusable in intrepid alpha 6...

any idea, where the issue is coming from?

Can you confirm that reverting back to 1:6.9.0+git20080802.1f3eee36-1ubuntu1 makes the issue go away?

Also, can you provide 'lspci -vvnn'? The generic lspci output doesn't provide the pci id's

Also, I noticed this in the log:

(II) RADEON(0): XAA Render acceleration unsupported on Radeon 9500/9700 and newer. Please use EXA instead.
(II) RADEON(0): Render acceleration disabled
(II) RADEON(0): num quad-pipes is 1
(II) RADEON(0): Using XFree86 Acceleration Architecture (XAA)

Does turning on EXA in xorg.conf make the issue go away? If so, perhaps this is a dupe of #267612

Another thing to try (mentioned in another dupe) is:

Option "DMAForXV" "off"

269357 may be another dupe of this issue. Sounds like we may need to have certain chipsets use EXA. 'lspci -vvnn' will be important for determining this.

To confirm the upstream fix, you may try the -ati packages in my PPA.

Option "AccelMethod" "EXA" pretty much solved the issue, but it's not perfect
sometimes there's some garbage in the windows, or in the place of minimized windows on the desktop
one of the videos i ran for testing also had some garbage on it, (different from the original bug) which disappeared after re-opening

After a little longer testing I have to say that EXA causes high Xorg usage and laggy desktop for me.
I've disabled it again for now. Tried Option "DMAForXV" "off", but that doesn't solve the original issue.

I've installed xserver-xorg-video-radeon_6.9.0+git20080930.c359c2a3-0ubuntu0tormod_i386.deb from Tormod's PPA.
Now xv is fine, and Xorg doesn't consume so high CPU.
Although "(II) RADEON(0): XAA Render acceleration unsupported on Radeon 9500/9700 and newer. Please use EXA instead." is still in xorg log.
Thank you for the help.

Tormod’s package fixed the problem for me as well.

Excellent, thanks. There's five changesets in that package, one of which is the fix for fdo #17254 which we already identified as the upstream of this bug. I'll cherrypick that fix.

I think I have this bug, too. I just upgraded to Interpid.
Using 1:6.9.0+git20080826.a3cc1d7a-2ubuntu3 (intrepid)
When I went from Hardy to Intrepid, I had to install the ATI driver. The flgr driver disappeared ??

When Intrepid first boots, the display flickers. It tries to display at 1900x1200 (my monitor's resolution)
Then in between flickers, I have to go into the K launcher and select the Display properties.
The resolution goes to a newer value and then everything is more or less ok.

The setting disappears after a reboot. I have to go through the process of resetting the display resolution.

hope this helps! tia!

lspci -vvnn

01:05.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon HD 3200 Graphics [1002:9610]
        Subsystem: Giga-byte Technology Device [1458:d000]
        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: 4 bytes
        Interrupt: pin A routed to IRQ 18
        Region 0: Memory at d0000000 (32-bit, prefetchable) [size=256M]
        Region 1: I/O ports at ee00 [size=256]
        Region 2: Memory at fdee0000 (32-bit, non-prefetchable) [size=64K]
        Region 5: Memory at fdd00000 (32-bit, non-prefetchable) [size=1M]
        Capabilities: [50] Power Management version 3
                Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 PME-Enable- DSel=0 DScale=0 PME-
        Capabilities: [a0] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
                Address: 0000000000000000 Data: 0000

Hi Johan,

I've built a package with the patch included which in theory should solve the issue. Would you mind testing and validating that it does indeed fix it?

http://www.bryceharrington.org/ubuntu/Ati/xserver-xorg-video-ati_6.9.0+git20080826.a3cc1d7a-2ubuntu4~bwh2_i386.deb

If you need the amd64 package or need the dbg symbols, those debs are at http://www.bryceharrington.org/ubuntu/Ati as well.

kelargo, from your description you have some bug other than this one. This bug only shows flickering when playing Xv. You should probably file that as a new bug, via `ubuntu-bug xserver-xorg-video-ati`.

And yes, -fglrx is not yet available for Intrepid users.

Tormods ...0930.deb solved it here... Your patched version Bryce still has the same issue...

By the way, I have this card:

01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon RV250 [Mobility FireGL 9000] [1002:4c66] (rev 01)
 Subsystem: Compaq Computer Corporation Device [0e11:0860]
 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: 128 (2000ns min), Cache Line Size: 32 bytes
 Interrupt: pin A routed to IRQ 10
 Region 0: Memory at 98000000 (32-bit, prefetchable) [size=128M]
 Region 1: I/O ports at 3000 [size=256]
 Region 2: Memory at 90400000 (32-bit, non-prefetchable) [size=64K]
 [virtual] Expansion ROM at 90420000 [disabled] [size=128K]
 Capabilities: <access denied>
 Kernel modules: radeonfb

ok, thanks. Bug #276903:

Hmm, that's rather surprising that the patched driver didn't solve the issue.

Well, here's another one that pulls in the atom bios changes as well. None of the other changes in the 0930 deb look like they'd matter for video issues.

http://bryceharrington.org/ubuntu/AtiBug267297/

Please test this deb and let me know if it solves the issue.

Johan, I'd also like to hear your results from testing http://www.bryceharrington.org/ubuntu/Ati/xserver-xorg-video-ati_6.9.0+git20080826.a3cc1d7a-2ubuntu4~bwh2_i386.deb as well.

Hi Bryce,

that package solves the issue. Just out of interest, why not take the upstream version here, i.e. as i've understood Tormod's deb? Thanks for your work on it.

Thanks for confirming those two patches fix it.

Well, I wanted to understand *what* changes fixed the bug, so I'm not just blindly copying a new upstream version and risking adding regressions for some people.

However, I understand all the other changes aside from the atom bios patch, and they seem safe. I'd hoped to cherrypick them, and avoid the atom bios changes (and thus risk any chance of regressions), but since we can't, I guess there's not much reason not to take a newer git version. While I can't prove it 100%, I trust the atom bios patch is safe and doubt we'll see much regression risk.

This bug was fixed in the package xserver-xorg-video-ati - 1:6.9.0+git20081003.f9826a56-0ubuntu1

---------------
xserver-xorg-video-ati (1:6.9.0+git20081003.f9826a56-0ubuntu1) intrepid; urgency=low

  * Pull upstream snapshot, up to commit f9826a56 (LP: #277104)
    - Add support for DDC via atom commands for RV410 (LP: #227520?)
    - man page updates (fdo #17887)
    - Make initial rotation work, by initializing acceleration _before_
      setting the mode
    - Fix Xv video corruption by passing base offset into
      RADEONDisplayVideo() explicitly (fdo #17254, LP: #267297)
    - Fix sync-to-vblank hangs after mode switch with DRM vblank-rework.
    - pll adjustments
    - Add quirk to fix macbook hardware (fdo #17625)
  * Drop 103_rotation_reset.patch - already upstream

 -- Bryce Harrington <email address hidden> Fri, 03 Oct 2008 18:56:44 -0700