Display corruption after short period

Bug #270617 reported by phenest
22
Affects Status Importance Assigned to Milestone
nvidia-graphics-drivers-177 (Ubuntu)
Invalid
Undecided
Unassigned

Bug Description

Ubuntu Intrepid
Linux precision 2.6.27-3-generic #1 SMP Wed Sep 10 16:18:52 UTC 2008 x86_64 GNU/Linux
nVidia driver 177.70

Not sure when or why it happens, but it happens with all nVidia drivers presently in Intrepid.

The only way out of this is to reboot. The computer is still working fine otherwise, and responds to input.

Revision history for this message
phenest (steve-clark) wrote :
Revision history for this message
phenest (steve-clark) wrote :

This appears to be similar to bug 109643

Revision history for this message
Alberto Milone (albertomilone) wrote :

Can you attach your /var/log/Xorg.0.log and post the model of your graphics card?

Revision history for this message
phenest (steve-clark) wrote :

Xorg.0.log attached

nVidia GeForce 7950GTX

Revision history for this message
phenest (steve-clark) wrote :

I have found that by running my laptop in a lower power state, that is setting the CPU freq. to powersave, this seems to help. Also, I'm using a basic stripped down version of xorg.conf which I'm letting nvidia-settings change only.

Revision history for this message
phenest (steve-clark) wrote :

If I run nvidia-settings in Intrepid, the GPU thermal temperature rises to 60C before the fan cuts in, and cools it down to 50C. In Hardy, the fans cut in at 50C, and keeps cooling until it reaches 40C. So it runs 10C cooler in Hardy.

Revision history for this message
Markus Golser (golserma) wrote :

I have this bug too (confirming)

Changed in nvidia-graphics-drivers-177:
status: New → Confirmed
Revision history for this message
Tuomas Aavikko (taavikko) wrote :

Bug #251448 Is sort of an similar case. Just tried to switch TTY1 and
(II) Video Bus: Device reopened after 10 attempts.
Should this be reopened, and mark as dublicate?

Revision history for this message
Markus Golser (golserma) wrote :

The bug does not look similar to me.
I just use the driver and wait.
He has to make actions in order to trigger the bug

Revision history for this message
phenest (steve-clark) wrote :

Incidentally, using driver 177.78 from nVidia never crashed with me. But 177.80 from nVidia does, almost immediately.

Revision history for this message
Markus Golser (golserma) wrote :

Windows Vista and Ubuntu Hardy works just fine on the same machine

I created a thread on the nvidia forums http://www.nvnews.net/vbulletin/showthread.php?p=1804424#post1804424

Revision history for this message
Markus Golser (golserma) wrote :

Just found out that i can trigger this bug using nvclock.
If I change the MHZ the system crashes exactly as earlier described.

Does the new driver or maybe the kernel change the clock speed somehow in order to save power?
If yes how can I dissable this?

Revision history for this message
phenest (steve-clark) wrote :

The only nVidia driver that is stable is 177.78. All others including 177.80 crash. Perhaps Markus could test 177.78 with nvclock.

Revision history for this message
Markus Golser (golserma) wrote :

177.78 has the same issues for me.
The screenshot is interesting

Revision history for this message
phenest (steve-clark) wrote :

Actually, 177.78 just crashed on me. Grrr! Very annoying.

Revision history for this message
Markus Golser (golserma) wrote :

I had to limit the GPU and the Memory to 100MHZ.
It is stable now.

modified xorg.conf attached

Revision history for this message
phenest (steve-clark) wrote :

I'm getting this with the nVidia beta 180.06 driver too. Almost tempted to change to ATI.

Revision history for this message
sklp (sklp) wrote :

I'm getting this on my laptop (Dell Inspiron 9400 -- GF Go 9600GS), which I have just bought. So the only OS I've tried is Intrepid (64-bit and 32-bit), and this issue appears for me sometimes and It's really annoying to have to reboot. I've got the bug with all 3 nvidia options in intrepid.
Markus Golser, I don't really want to try your fix as I would like my GPU to perform when I'm playing games etc...
I'm wondering if this issue does not appear at all in hardy, and I'm willing to downgrade for a while if it's stable. Also I'm interested if there's any tricks you want to share.
phenest, in the the 109643 bug you mention that it gets fixed by tuning the cpu powersave to "performance", is this actually/still the case? I may want to try that although it seems lame to have a laptop without speedstep capability :/
Downgrading to hardy may be the best option...

Revision history for this message
sklp (sklp) wrote :

CORRECTION: I meant GeForce Go 7900GS in my last post.

Revision history for this message
phenest (steve-clark) wrote :

The only fix that works is the one given by Markus. Otherwise, yes, go back to Hardy.

Revision history for this message
Markus Golser (golserma) wrote : Re: [Bug 270617] Re: Display corruption after short period

Someone knows if upstream (nvidia) knows about this bug?

This is a mayor bug if you ask me...

Revision history for this message
sklp (sklp) wrote :

I used hardy for a while, without any problems. However, today I got the exact same problem I got in intrepid(looks kinda lika the image above). So the problem apparently exists in hardy as well, but it just seemed to take longer for the problem to appear (for me, but it might just be random). Guess I'll have to use Windows for a while on this machine, sadly.

Revision history for this message
sklp (sklp) wrote :

Anyone tried 180.11 (perhaps in jaunty)?

Revision history for this message
Markus Golser (golserma) wrote :

I'm currently testing 180.11...
It did not crash so far... I report later if it works on the long run

Revision history for this message
Markus Golser (golserma) wrote :

Reproduced on Intrepid 32bit using 180.11
Where is the love for Nvidia users???

I'm back to
        Option "Coolbits" "1"
        Option "RegistryDwords" "PowerMizerLevelAC=0x3"
        Option "RegistryDwords" "PowerMizerLevel=0x3"

Revision history for this message
phenest (steve-clark) wrote :

It happened to me within 5 minutes. This is getting infuriating. My card and driver work perfectly under Hardy, but have not worked right since. I wish I could change to a different brand, but I'm stuck with nVidia with this laptop. Hmmph!

Revision history for this message
Markus Golser (golserma) wrote :

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

@phenest did you try to use the hardy kernel on intrepid?
Maybe using an older kernel works
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (MingW32)
Comment: http://getfiregpg.org

iEYEARECAAYFAklGcwMACgkQayJvJgkUaGL43ACff/p1heT8Y80XHIBg7HuSSgH3
p5oAnjIqi++0uZaB4tHGG2Gqdwd1iPty
=0GAH
-----END PGP SIGNATURE-----

Revision history for this message
Markus Golser (golserma) wrote :

Have installed 2.6.28-rc8 on Intrepid... 4 hours 0 crashes...

Revision history for this message
Markus Golser (golserma) wrote :

Reproduced using 2.6.28-rc8 on Intrepid :(

Revision history for this message
sklp (sklp) wrote :

In my last post i described getting the bug in hardy. That occured with the latest stable driver (from hardy repos).
I've since then installed the beta driver(180.11) in hardy, and It's been stable for many days. I can't be certain it's 100% stable, but it certainly seems so at the moment.
However, it may not be stable in intrepid or jaunty (you seem to have problems with 180.11 and intrepid), but I'll probably try jaunty in a while, as intrepid seems to be unstable with every driver version I've tried so far.

Revision history for this message
sklp (sklp) wrote :

Haven't had a single crash since my last post (with the same setup)

Revision history for this message
Markus Golser (golserma) wrote :

sklp: You have the Bug on Intrepid and not on Hardy right?

Revision history for this message
sklp (sklp) wrote :

Yes I have the bug in intrepid with every driver version I've tried so far. However, I also discovered the bug once in hardy with the latest driver from repositories. But haven't seen it yet after a long time of testing (using 180.11 beta on hardy 64-bit). So I'm not switching driver and/or ubuntu version anytime soon. Maybe when jaunty is released I'll try that.

Revision history for this message
Mark Mandel (mark-mandel) wrote :

Not sure if it helps at all, but there is a 3 page discussion of this bug at:
http://ubuntuforums.org/showthread.php?t=1006396

With screen grabs, and I believe some logs as well.

I've had this occur for me a well, on every 177.x version of the nVidia driver that I have run. Currently running 177.82.

The really weird thing is - it NEVER happens if I have a 2nd monitor attached, and I'm in twinview. It only every happens to me when I'm displaying on a single screen on my laptop.

Dell XPS M1710
NVidia GeForce GO 7950 GTX Card

Revision history for this message
stoneybridge (stoney-tinyonline) wrote :

daft question mark, hows the 180.22 drivers for you? seems we have the same lappy, my cpu is a core duo t2600 2.16
ive 3 gig of ram and the same gfx card (512meg) running 8.10, as ive said in the other forum you've linked to, i had this prob but now it's sorted,

maybe i could be of sum help, if you don't get too technical i've only been playing with linux for a month or so

Revision history for this message
phenest (steve-clark) wrote :

My laptop is a Dell Precision M90 which is actually an XPS M1710 but designed for the business market (aesthetic changes). It normally ships with an nVidia Quadro but I upgraded it to an nVidia Go 7950 GTX. So it looks like we all use the same laptop and the same video card. I haven't tried your fix stoneybridge, but I'll take a look.

Revision history for this message
phenest (steve-clark) wrote :

I just installed 180.22, set the resolution, refresh rate, and Video Blitter Adapter Settings is unchecked. I'll keep you informed of anything. Nice to know that closing and opening the lid restores the screen. Have you tried this too Markus?

Revision history for this message
Markus Golser (golserma) wrote :

closing and opening the lid does not work for me

Revision history for this message
phenest (steve-clark) wrote :

It just crashed on me. Closing and opening the lid does work if you have it set to go blank when closed (Power Manager). I have posted a comment on nvnews http://www.nvnews.net/vbulletin/showpost.php?p=1923143&postcount=18. I'm now going to repotr this to nVidia. To date, only the fix suggested by Markus has any lasting effect.

Revision history for this message
phenest (steve-clark) wrote :

Here is a link if anyone wants to report to nVidia directly: http://nvidia.custhelp.com/cgi-bin/nvidia.cfg/php/enduser/std_adp.php?p_faqid=44

Revision history for this message
phenest (steve-clark) wrote :

As this bug is not Ubuntu specific, I have marked it as Invalid. If you wish to continue the discussion, please follow the thread on the Ubuntu Forums: http://ubuntuforums.org/showthread.php?p=6677361#post6677361 or at nvnews: http://www.nvnews.net/vbulletin/showthread.php?p=1923212&posted=1#post1923212

Changed in nvidia-graphics-drivers-177:
status: Confirmed → Invalid
Revision history for this message
Markus Golser (golserma) wrote :

I reported this bug 2 months ago to nvidia. No response so far! I guess they don't care.

This bug is an Ubuntu regression everything worked just fine on Hardy

Revision history for this message
phenest (steve-clark) wrote :

This issue is being reported on all the other distros too. Not just Ubuntu

Revision history for this message
Markus Golser (golserma) wrote :

Same problem using 180.29

Revision history for this message
sklp (sklp) wrote :

I have used 180.xx on Hardy 64 bit without issues for a while... I recently upgraded to jaunty though, and it's been fine for a couple of days but now i got the bug.
However, i was able to get rid of it by doing some vt switching back and forth :)

Revision history for this message
Markus Golser (golserma) wrote :

Looks like this is a UBUNTU problem...

I switched to the mainline kernel 2.6.27-02062719-generic and nvidia 180.29 (on intrepid) and did not have this bug for 5 hours....

Revision history for this message
Markus Golser (golserma) wrote :

now I had the bug just after posting this :(
Lets see if 2.6.29-020629rc7 works

Revision history for this message
Markus Golser (golserma) wrote :

I'm on archlinux now.

X.Org X Server 1.5.3
Linux 2.6.28
nvidia 180.29

->> no problems at all

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Duplicates of this bug

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.