[RS690M X1200] (Needs 55fbdbae) screen randomly blanks for short times

[David]

Binary package hint: xserver-xorg-video-radeon

I'm using Ubuntu 8.10 with the open source Radeon driver (used by default after installation). By default the driver doesn't detect my display's size and runs at 1024x768. In this mode everything works fine. If I add 'Option "PanelSize" "1680x1050"' to my xorg.conf and restart, the driver uses the display's correct native resolution.

At the higher resolution, the screen now randomly blanks out for a second. On an idle system these blanks are rather seldom, mabye once per 5 minutes. If I enter a 3D game, such as ppracer, these blanks occur much more frequently, many times per minute.

This looks like the 'display buffer underflow' problem described in 'man radeon'. Unfortunately setting option "DisplayPriority" to "HIGH" or "BIOS" doesn't help.

$ lspci -d 1002:791f
01:05.0 VGA compatible controller: ATI Technologies Inc RS690M [Radeon X1200 Series]

snipset from 'apt-cache policy xserver-xorg-video-radeon':
Installed: 1:6.9.0+git20081003.f9826a56-0ubuntu2

I'm using a 1680x1050 display connected digitally via my board's HDMI output. I only have a single 2GB SO-DIMM DDR2-800 RAM module on my mainboard, which is used for the shared video RAM.

Copy of Xorg.0.log attached.

Using the proprietary fglrx driver, I don't see those blanks, which should rule out a hardware problem.

[goto (gotolaunchpad)]

Does the problem also appear if you use the radeonhd driver instead of the driver "radeon"?
Thanks.

[David (dvdkhlng)]

Yes, same problem with radeonhd.

[Daniel Newman (dmnewman)]

I'm having the same random blanking problem on a HD3450 at 19200x1200 driven through the HDMI output.

xserver-xorg-video-radeon:
  Installed: 1:6.9.0.91-1ubuntu4

[goto (gotolaunchpad)]

I think it would be wise to report this upstream.
Look here, how this is to be done:
https://wiki.ubuntu.com/Bugs/Upstream/Freedesktop

Thanks.

[Daniel Newman (dmnewman)]

Further information. The blanking problem also occurred using the DVI output. However, it does not occur on either HDMI or DVI using this earlier version,

xserver-xorg-video-radeon:
  Installed: 1:6.9.0+git20081003.f9826a56-0ubuntu5

Also, I've just noticed, in my previous post, I really meant 1920x1200 :)

[Daniel Newman (dmnewman)]

still happening on the HD3450 with version

1:6.10.0-1ubuntu1

It seems to happen more frequently when the display content is changing rapidly.

[goto (gotolaunchpad)]

Daniel, does the problem exists for you, too if you use the 'radeonhd' driver instead of 'radeon'?
There is no other way than filing a bug report upstream :D
I can do that for you if you don't want to do that, but you have to provide logs and so on if it's necessary.
Shall I file the bug report for you?

[Daniel Newman (dmnewman)]

The radeonhd driver doesn't work at all for me - it goes through the ubuntu splash screen and then remains blank. I have attached the logs for both the radeon and radeonhd sessions, and the xorg.conf used in both cases. Please file the report wherever necessary.

[goto (gotolaunchpad)]

Daniel: Please test also if you can get radeonhd to work with the option described here:
http://www.x.org/wiki/radeonhd
(in section "10.5. My monitor isn't detected").

If no option helps, I'll file a bug report upstream.
Thanks!

[Daniel Newman (dmnewman)]

Tested several times using radeonhd driver with option HPD set to "swap"
and "off". Did not work in either case. The log for HPD off is
attached. Both cases gave a black screen, sometimes with a few pixels
along the top of the screen in white, red or blue.

goto wrote:
> Daniel: Please test also if you can get radeonhd to work with the option described here:
> http://www.x.org/wiki/radeonhd
> (in section "10.5. My monitor isn't detected").
>
> If no option helps, I'll file a bug report upstream.
> Thanks!
>
>

[goto (gotolaunchpad)]

Daniel,
so radeonhd gives you no picture at all. Could you please remove the HPD option from your xorg.conf again and then start x with:
startx -- -logverbose 7
And then please load up the xorg.0.log.
I need just a bit more information: Which brand has you radeon HD3450, and which monitor do you try to connect over which connection (VGA/DVI/HDMI...)?
Then I'll forward your bug report, thanks! :)

[Daniel Newman (dmnewman)]

Ok, the attached xorg.conf and log are as requested. I have also
included lspci -vvv for the video card. The card itself is a HIS HD
4350 (see www.hisdigital.com) 64bit card with onboard fan.

These files were generated using radeonhd, trying to connect over DVI to
an ASUS MK241H lcd monitor, which has a native resolution of 1920x1200.

Please let me know if you need any more information

goto wrote:
> Daniel,
> so radeonhd gives you no picture at all. Could you please remove the HPD option from your xorg.conf again and then start x with:
> startx -- -logverbose 7
> And then please load up the xorg.0.log.
> I need just a bit more information: Which brand has you radeon HD3450, and which monitor do you try to connect over which connection (VGA/DVI/HDMI...)?
> Then I'll forward your bug report, thanks! :)
>
>

[goto (gotolaunchpad)]

Daniel: Done. The bug is located at
https://bugs.freedesktop.org/show_bug.cgi?id=19700
I'll report back if the devs want more information.
Thanks for your help. :)

[Bryce Harrington (bryce)]

goto, thanks for helping Daniel report the problem running -radeonhd, however he is not the original reporter, and the original reporter's bug was a different issue - short infrequent blanks seen on the -ati driver. So that upstream bug sort of takes this bug report off track. I'll remove the bug link so we can focus on the original reporter's original issue.

[Bryce Harrington (bryce)]

Oh, actually the bug link is the appropriate bug. We can focus on that one.

[Martin Emrich (emme)]

Hi!

I have the same problem here. I am running jaunty amd64 on a Gigabyte GA-MA69GM-S2H mainboard (AMD 690G). With xserver-xorg-video-ati (6.11.0+git20090310.945ccbbd-0ubuntu1) I get these random screen blanks on my DVI output (the second VGA head is unaffected). Here, too, the blanks occur more often and are longer if the system is under higher load. 'Option "DisplayPriority" "HIGH"' did have no effect.

I just gave radeonhd a try, and I had no blanks for the last 20 minutes. But with it I don't have XVideo or DRI, and even the 2D performance is very low.

[goto (gotolaunchpad)]

Martin: With radeonhd you would need to enable DRI by hand ind your xorg.conf. But there is a good reason, that it's not enabled automatically, it is not 100% stable yet. (In the version in Ubuntu at least).

[teta (velten)]

Hello

Same problem here after update to jaunty. (screen randomly blanks for short times)
System: amd64
lspci gives:
01:00.0 VGA compatible controller: ATI Technologies Inc RV516 [Radeon X1300/X1550 Series]
01:00.1 Display controller: ATI Technologies Inc RV516 [Radeon X1300/X1550 Series] (Secondary)

I have no Problem with the radeonhd-Driver:

I removed Packages: xserver-xorg-video-ati, xserver-xorg-video-all
radeonhd Package installed: xserver-xorg-video-radeonhd
Old xorg.conf changed with: dpkg-reconfigure xserver-xorg

works for me.

[Bernhard Polzin (b-polzin)]

Hi!

For completeness:
I got the problem by update from 8.10 to 9.04. The screen became black for less than a second, randomly. No relation to graphics or cpu load identifiable.

System: Intel 32bit
CPU: Intel(R) Pentium(R) 4 CPU 3.20GHz stepping 01
Card: ATI Sapphire X1550, 256MB
Monitor: NEC Multisync LCD2690WUXi with resolution 1920x1200 at 60Hz connected via DVI

lspci:
01:00.0 VGA compatible controller: ATI Technologies Inc RV516 [Radeon X1300/X1550 Series]
01:00.1 Display controller: ATI Technologies Inc RV516 [Radeon X1300/X1550 Series] (Secondary)

Since installation of package xserver-video-radeonhd and restart of the machine everything is fine, but it seems to be slower (in showing videos).

Hope it helps

[Tom-ipe-systems (tom-ipe-systems)]

I have had some problems with flickering using this chip and high resolution displays over DVI/HDMI. The problem seems dependent on the model of screen, some seem more sensitive to variations in pixel clock than others. I have managed to resolve the issue by making 2 changes to radeon_crtc.c, one to enable the use of the fractional fb divider and secondly by changing the selection criteria for PLL values to choose the closest frequency that is less than the target frequency, rather than just the closest. I have added these changes unconditionally to the driver (it's basically a quick hack to make it work and certainly could be much tidier), but can post a patch if required.

I don't have access to any other radeon cards so I can't say if this problem or the solution affects them - I suspect from the looking at the commit logs that the fractional fb div doesn't work on some cards, but it seems to be required for the RS690.

Any advice on how/where to properly incorporate these changes into the driver would be appreciated.

[agd5f (agd5f)]

Please attach the patch and your xorg log.

[Tom-ipe-systems (tom-ipe-systems)]

Created an attachment (id=26677)
RS690 PLL generation

[Tom-ipe-systems (tom-ipe-systems)]

Created an attachment (id=26685)
Xorg Log with 1920x1200 DELL screen - pre patch

[Tom-ipe-systems (tom-ipe-systems)]

Created an attachment (id=26686)
Xorg Log with 1920x1200 DELL screen - patch applied

[Tom-ipe-systems (tom-ipe-systems)]

Created an attachment (id=26687)
Xorg Log with 1080p screen - fract fb div enabled

This log is with on a 1080p screen with the factional fb div enabled, but without the second change, and produces a flicker or sometimes just random pixel corruption (similar to the snow effect pictured in bug #21553), hence the need for selecting the closest PLL value below the desired frequency.

[Tom-ipe-systems (tom-ipe-systems)]

Created an attachment (id=26688)
Xorg Log with 1080p screen - patch applied

[agd5f (agd5f)]

Thanks Tom, I've gone ahead and incorporated the patch. It seems to work fine on all the hardware I have access to.

[agd5f (agd5f)]

*** Bug 15175 has been marked as a duplicate of this bug. ***

[Tom-ipe-systems (tom-ipe-systems)]

Alex - I'm sure you are aware of this, but this fix is also relevant for the radeon KMS code in Fedora 11 and submitted for 2.6.31, would be good to see it applied there too if possible.

[agd5f (agd5f)]

(In reply to comment #9)
> Alex - I'm sure you are aware of this, but this fix is also relevant for the
> radeon KMS code in Fedora 11 and submitted for 2.6.31, would be good to see it
> applied there too if possible.
>

Yes, I plan to queue it up along with any other ddx fixes that have accumulated since the my last push form the ddx to kms.

[Vincenzo Ciancia (vincenzo-ml)]

I have this bug on jaunty why was this marked as invalid?

[Vincenzo Ciancia (vincenzo-ml)]

Sorry, it's only upstream. BTW it seems fix released upstream, not invalid.

[Martin Emrich (emme)]

The upstream bug was marked as a duplicate. http://bugs.freedesktop.org/show_bug.cgi?id=20222 is related to this one, and still open.

[Tormod Volden (tormodvolden)]

Please try the -ati/-radeon packages for Jaunty in https://launchpad.net/~tormodvolden/+archive/ppa to see if the upstream fix helps here.

[Martin Emrich (emme)]

The bug is still present in upstream's git. I am currently trying to help the upstream developers to find the cause...

Anyone also experiencing this problem: How is your affected display connected? What's the native resolution of your affected display? What's the actual configured resolution? Does the flickering go away if you select a lower resolution?

[Vincenzo Ciancia (vincenzo-ml)]

Will reply in two weeks properly, because I will not be at office very soon and the machine is there. I use a digital monitor with a high resolution connected via vga. Not so frequently (can be once per hour) the monitor goes black and comes back after some seconds.

[Ian Pilcher (arequipeno)]

FYI, this patch seems to fix the occasional "flicker" that I was seeing on my
X1650 with 2 DVI-attached LCDS (1680x1050 & 1280x1024). Nice work!

[Bryce Harrington (bryce)]

This is the patch of the upstream commit that purports to fix the bug, however it isn't applying to the -ati we have in Karmic now, so we'll pick it up after we update to a newer -ati.

[David (dvdkhlng)]

After upgrading to Jaunty (9.04), the problem persists. Even more troubling now, since starting with Jaunty my graphic chip (RS690M) is not supported by the proprietary fglrx driver any more :(

Current workaround: after logging in, switch once to the console screen (Ctrl+Alt+F1), then switch back to the Xserver (Ctrl+Alt+F7). After that procedure, the blanking is gone, but has to be repeated whenever I log in.

[Launchpad Janitor (janitor)]

This bug was fixed in the package xserver-xorg-video-ati - 1:6.12.99+git20090825.fc74e119-0ubuntu1

---------------
xserver-xorg-video-ati (1:6.12.99+git20090825.fc74e119-0ubuntu1) karmic; urgency=low

  * Checkout from git 20090825 (master branch) up to commit
    fc74e1194c980d978667e02c60a29a761a694bde
    + Adds DRI2 / redirected direct rendering
      (LP: #96991)
    + Fix freeze on opengl games (bzflag, alienarena) (when KMS on)
      (LP: #348450)
    + Fix intermittent short screen blanking behavior
      (LP: #310864)
    + Fix faulty wine screen updates with EXA Compositing
      (LP: #314205)
    + Fix screen corruption issues when switching windows
      (LP: #406731)
    + Fix SIGSEGV crash in drmCommandNone()
      (LP: #352567)
    + Fix smeared/tearing of display on reboot
      (LP: #367741

  [Tormod Volden]
  * 199_add_git_version_to_log.diff: Log git commit id in RadeonPreInit()
  * sed -i s/DRI2BufferPtr/DRI2Buffer2Ptr/ src/radeon_dri2.c
      so it compiles against xserver 1.6.2.+

  [Bryce Harrington]
  * Restore patch system and quilt to build-depends
  * Drop patches already present upstream
    + 107_check_unsupported_composit_ops.patch
    + 108_quirk_agpmode_m6_ali.patch: AGPMode quirk.
    + 109_quirk_agpmode_m7_intel.patch: AGPMode quirk.
    + 110_quirk_agpmode_r420_sis.patch: AGPMode quirk.
    + 111_use_xaa_for_lowmem_or_nodri.patch

 -- Bryce Harrington <email address hidden> Sat, 29 Aug 2009 12:16:53 -0700

[Martin Emrich (emme)]

Good to hear that! As I currently use an X800 GTO, I'll wait for the weekend to test the new package with the onboard RS690.

[Martin Emrich (emme)]

Just found some free minutes, and pulled out the X800. Sadly, the blackouts are still there. Note that Davids workaround (VT-switching) has no effect here.

[Martin Emrich (emme)]

This is the freedesktop.org bug that I personally associate with this one: http://bugs.freedesktop.org/show_bug.cgi?id=20222
If you wish, I can file a separate new bug against Ubuntu's xserver-xorg-video-ati.

[Shaun Crampton (fasaxc)]

I just tried the Karmic beta and I'm still seeing this bug with my Radeon HD3300, 1900x1200 monitor connected by DVI. Every few minutes my screen blanks for ~1second. Should it be fixed in the beta? If so, I don't think the fix has worked.

[Bryce Harrington (bryce)]

The title of this bug is unfortunately generic, however it appears there are several different underlying bugs in the system which produce similar symptoms (random screen blanks). The particular bug investigated upstream for this report was fixed by the referenced patch.

We maintain bug reports on a per-bug basis rather than a per-symptom basis, so since the particular bug investigated for this bug report was solved, this report should be closed, even if people are still seeing the same symptom (due to a different issue). I know that sounds confusing and perhaps frustrating, but this workflow is necessary to make solving these problems feasible. Otherwise we end up with one bug report that is a mishmash of different discrete bugs and it becomes ambiguous what their status is; this greatly increases the difficulty to developers working on the bug, which isn't in our interest. ;-)

Bug #331522 is tracking one remaining screen blanking bug, so I would encourage you to watch that one. If you are having this symptom and believe it is due to a bug different from #331522, please file a new bug.

[Basil Baumann (basil-baumann)]

I am not very familiar to linux, but I have the same problem discribed above with an EN9600GT and an ASUS MK241H Screen.
Is the described patch also working with my equipment?
Is there some description, how to install this patch?
Thanks!

[Natim (site-remy)]

Same problem here on Ubuntu 11.04 with :

01:05.0 VGA compatible controller: ATI Technologies Inc RS690M [Radeon X1200 Series]

Unity lauches but we get some glinches that make the computer unusable.