Bug 19769 - [Q45/G45] green(!) pixel flicker with Intel DQ45EK/DG45FC boards using DVI output
Summary: [Q45/G45] green(!) pixel flicker with Intel DQ45EK/DG45FC boards using DVI ou...
Status: RESOLVED WORKSFORME
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/intel (show other bugs)
Version: 7.4 (2008.09)
Hardware: x86 (IA32) Linux (All)
: medium normal
Assignee: MaLing
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords: NEEDINFO
Depends on:
Blocks:
 
Reported: 2009-01-27 14:06 UTC by Hans-Peter Jansen
Modified: 2009-02-18 23:50 UTC (History)
7 users (show)

See Also:
i915 platform:
i915 features:


Attachments
xorg.conf (4.85 KB, text/plain)
2009-01-27 14:06 UTC, Hans-Peter Jansen
no flags Details
Xorg.0.log (29.68 KB, text/x-log)
2009-01-27 14:07 UTC, Hans-Peter Jansen
no flags Details
hwinfo --gfxcard (1.60 KB, text/plain)
2009-01-27 14:07 UTC, Hans-Peter Jansen
no flags Details
hwinfo --bios (10.18 KB, text/plain)
2009-01-27 14:07 UTC, Hans-Peter Jansen
no flags Details
detail photo (63.95 KB, image/jpeg)
2009-01-27 14:08 UTC, Hans-Peter Jansen
no flags Details
detail photo (57.05 KB, image/jpeg)
2009-01-27 14:08 UTC, Hans-Peter Jansen
no flags Details
detail photo (62.78 KB, image/jpeg)
2009-01-27 14:08 UTC, Hans-Peter Jansen
no flags Details
"funny" screen background (143.32 KB, image/jpeg)
2009-01-27 14:09 UTC, Hans-Peter Jansen
no flags Details
Xorg.0.log with ModeDebug option enabled (74.73 KB, text/x-log)
2009-02-03 15:05 UTC, Hans-Peter Jansen
no flags Details
the issue doesn't occur on DG45FC (76.97 KB, text/plain)
2009-02-12 01:44 UTC, MaLing
no flags Details

Description Hans-Peter Jansen 2009-01-27 14:06:36 UTC
Created attachment 22282 [details]
xorg.conf

System: openSUSE 11.1/i386

I suffer from green pixel flicker with the new Intel DQ45EK and DG45FC board, but will report the details of the former system, as I stored the other for later investigation (I'm space and time contrained at the moment, catched the flu, etc.., please bear with me). Most probably fixing one of them may fix it for the other magically.

The symptoms are flickering wrong colored pixels of certain ranges, thus the
openSUSE (KDE4) kdm and desktop look pretty strange: along the green background pattern, there are flickering zones (see attached pictures).

After switching to VGA via an adapter (the board provides two DVI ports only) and creating a new config with sax2, the pixel flicker vanished.
Well, the display is a bit blurred (as expected), and when switching screens,
funny colors appear (e.g. during X startup or shutdown, screenshot attached).

Switching back to DVI resulted in reappearing flicker (looks like the flicker
happens with some values of blue also (a few pixel in sax2 test screen do it as
well..). 

The other board doesn't even provide an DVI to VGA adapter, and generally -
using VGA connects is a bit retro, isn't it ;-) Anyway, I planned to connect to
a 24" (1920x1200) screen, where VGA isn't an 'real' option anymore if at all.

I formerly reported this issue here, but it was identified as a hard to fix issue, and I was requested from Stefan Dirsch to file the report here.

https://bugzilla.novell.com/show_bug.cgi?id=469243
Comment 1 Hans-Peter Jansen 2009-01-27 14:07:01 UTC
Created attachment 22283 [details]
Xorg.0.log
Comment 2 Hans-Peter Jansen 2009-01-27 14:07:29 UTC
Created attachment 22284 [details]
hwinfo --gfxcard
Comment 3 Hans-Peter Jansen 2009-01-27 14:07:44 UTC
Created attachment 22285 [details]
hwinfo --bios
Comment 4 Hans-Peter Jansen 2009-01-27 14:08:12 UTC
Created attachment 22286 [details]
detail photo
Comment 5 Hans-Peter Jansen 2009-01-27 14:08:29 UTC
Created attachment 22287 [details]
detail photo
Comment 6 Hans-Peter Jansen 2009-01-27 14:08:50 UTC
Created attachment 22288 [details]
detail photo
Comment 7 Hans-Peter Jansen 2009-01-27 14:09:17 UTC
Created attachment 22289 [details]
"funny" screen background
Comment 8 Gordon Jin 2009-01-28 21:05:23 UTC
Note we have a DG45FC. Ling, the machine name is x-g45a.
Comment 9 Hans-Peter Jansen 2009-01-29 15:05:16 UTC
Gordon, it happens reproducible for me with both systems.

It would be nice, if you could confirm, that you're able to reproduce it, too.

Thanks in advance.
Comment 10 Hans-Peter Jansen 2009-01-29 15:07:16 UTC
BTW, if not, let me know, too. I would swap my machines and recreate the attached info.
Comment 11 Hans-Peter Jansen 2009-01-29 15:17:24 UTC
Ling, sorry for the confusion, I meant you for sure. 
Working for 16 hours is a bit too much with a flu :-[..

Let me know, if I can provide anything to help you. 

Matthias Hopf wrote something interesting (for you) in my original report:

http://bugzilla.novell.com/show_bug.cgi?id=469243#c18

Sorry again,
Pete

Comment 12 Michael Fu 2009-02-03 01:32:19 UTC
Is this HDMI specific? Why the title says HDMI? What if you connect to a DVI monitor?

I would also suggest to get rid of the conf file generated by sax and re-test...

thanks.
Comment 13 Hans-Peter Jansen 2009-02-03 04:35:09 UTC
> Is this HDMI specific? Why the title says HDMI? What if you connect to a DVI
> monitor?

Well, that was suggested by Stefan. In fact, I don't have any further HDMI hardware, and I connected the 19" LCD via a DVI cable. As noted before, using the DVI-VGA supplied adapter and connecting via VGA let the problem vanish (but raises others: blurred screen, etc..) Note: targeted screen/res is 24" and 1920x1200x24.

> I would also suggest to get rid of the conf file generated by sax and
> re-test...

Did something similar already: thrown out the HDMI related options, reduced the vert refresh rate (just in case..), fine tuned the mode lines, but the problem persists (since the issue is DVI related, I know fiddling with analog knobs isn't all that senseful, anyway)... After all, it seems like a timing issue.

Comment 14 Michael Fu 2009-02-03 04:59:06 UTC
ok. please add Option "ModeDebug" "True" to your device section , then attach the log. see http://intellinuxgraphics.org/how_to_report_bug.html

How many monitors have you tested? Have you tried a smaller one? What's the model of your monitor?


>The symptoms are flickering wrong colored pixels of certain ranges, thus the
>openSUSE (KDE4) kdm and desktop look pretty strange: along the green background
>pattern, there are flickering zones (see attached pictures).

flickering zones? this is probably hard to describe, but when we say flickering, it usually means the whole screen flickers. Do you see only a section of the screen flick(splash)? I'm asking because the sreenshot is static, I'm trying to understand more...


>I formerly reported this issue here, but it was identified as a hard to fix
issue,

Do you remember the bug number? Is this problem a regression for you, i.e. did it work before with the same set of HW?

thanks.
Comment 15 Hans-Peter Jansen 2009-02-03 15:04:25 UTC
> ok. please add Option "ModeDebug" "True" to your device section , then attach
> the log. see http://intellinuxgraphics.org/how_to_report_bug.html

Attached. I used this document to report this bug, anything missing?

> How many monitors have you tested? Have you tried a smaller one? What's the
> model of your monitor?

One, a VisionMagic T19DS (it's my first 19" monitor).

> flickering zones? this is probably hard to describe, but when we say
> flickering, it usually means the whole screen flickers. Do you see only a
> section of the screen flick(splash)? I'm asking because the sreenshot is
> static, I'm trying to understand more...

First, I'm not a native speaker. It's flickering pixels. Look at the first photo: left is the monitor frame (and some light reflection), down is part of the kde4 menu, the area between is normally smooth green forms, but note the pixel stripes: looking like dirt spots colored from dark grey and light green to orange, those areas are flickering (fast blinking in non deterministic order), and all these are not to be expected. 

>> I formerly reported this issue here, but it was identified as a hard to fix
>> issue,

Sorry, I confused this sentence, read:

>> I formerly reported this issue there
                                  ^

> Do you remember the bug number? Is this problem a regression for you, i.e. 
> did it work before with the same set of HW?

No, since these boards are pretty new, I started experimenting with them targeted as my next secondary office and wife compatible living room system.
openSUSE 11.1 is the first distribution, I've used with them, and would be happy to get them to work soon (some customers, I supply with diskless office systems have also signalled interest..).
Comment 16 Hans-Peter Jansen 2009-02-03 15:05:38 UTC
Created attachment 22551 [details]
Xorg.0.log with ModeDebug option enabled
Comment 17 Stefan Dirsch 2009-02-03 19:26:46 UTC
What happened to the summary? Reverting (unintended?) change.
Comment 18 Stefan Dirsch 2009-02-03 19:31:25 UTC
Michael, sure you did read what Matthias wrote in

  https://bugzilla.novell.com/show_bug.cgi?id=469243#c18

? He's also in Cc of this bugreport, in case you have any more questions.
Comment 19 Michael Fu 2009-02-11 19:37:13 UTC
(In reply to comment #15)
> 
> > flickering zones? this is probably hard to describe, but when we say
> > flickering, it usually means the whole screen flickers. Do you see only a
> > section of the screen flick(splash)? I'm asking because the sreenshot is
> > static, I'm trying to understand more...
> 
> First, I'm not a native speaker. It's flickering pixels. Look at the first
> photo: left is the monitor frame (and some light reflection), down is part of
> the kde4 menu, the area between is normally smooth green forms, but note the
> pixel stripes: looking like dirt spots colored from dark grey and light green
> to orange, those areas are flickering (fast blinking in non deterministic
> order), and all these are not to be expected. 
> 

ok. I'm kind of getting the idea.. What if you drag a window with white background on top of that area, will the pixels become normal(white) or still bad(random color?)? Just want to know if it's the background specifc or other pattern...
Comment 20 MaLing 2009-02-12 01:44:28 UTC
Created attachment 22853 [details]
the issue doesn't occur on DG45FC 

hi  Hans,
Today I did experiment on DG45FC, but no issue as your description occurs.
And I use Fedora, instead of openSUSE, could you try our latest version?

Thanks
Ma Ling
Comment 21 Michael Fu 2009-02-12 05:26:07 UTC
(In reply to comment #20)
> Created an attachment (id=22853) [details]
> the issue doesn't occur on DG45FC 
> 
> hi  Hans,
> Today I did experiment on DG45FC, but no issue as your description occurs.
> And I use Fedora, instead of openSUSE, could you try our latest version?
> 
> Thanks
> Ma Ling
> 

Very likely this issue is related with the monitor, I think. I.e. just we can't reproduce it may just because we don't have the HW. 

Jansen, do you have another DVI monitor to try?

Comment 22 MaLing 2009-02-18 17:44:02 UTC
ping Hans ~

Thanks
Ma Ling
Comment 23 Hans-Peter Jansen 2009-02-18 23:50:10 UTC
Sorry guys, I'm in the middle of a big relocation (company and family in one go.. 10% done), but yes, from a quick test, the issue didn't happen. Unfortunately other issues appeared, which may be related to the Eizo 30" screen, I've attached, but I guess, the HW just isn't capable of handling 2560x1600x24 natively (something I kind of accept, since both parts are targeted to very different places of action). Due to this issue, I will get rid of the old monitor soon...

Until I'm able to retest this issue with different HW, I've closed this issue with WORKSFORME. 

Sorry again for the churn, guys!


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.