Black dots appear on screen

[ruru]

On logging in to Ubuntu, sometimes there are randomly placed black dots on the screen. I have seen this problem on an old Hardy install on a PPC and on my current Intrepid install running on an Intel Q9550. Another machine with Intrepid on an Athlon has never exhibited this problem. All were 64-bit installs.

The dots are black pixels and not window or virtual desktop dependent (they remain in the same place regardless of what the OS draws on screen). Occasionally some may appear or disappear and sometimes they all go away and return at random times.

I have discovered that simply logging out of the session and logging back in to X clears this problem (I have had many weeks of uptime after this, so am certain that it is a permanent fix).

The dots do not affect any function of the OS, and can be fixed (for me) by restarting X (as above), but they are an unnecessary ugliness for those users that are affected.

My guess is that the issue is with the X server - but I am not sure how to verify this.

There is a thread (784045) about this on the Ubuntu forums (http://ubuntu-ky.ubuntuforums.org/showthread.php?t=784045)

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation 4 Series Chipset DRAM Controller [8086:2e20] (rev 02)
     Subsystem: ASUSTeK Computer Inc. Device [1043:82d3]
01:00.0 VGA compatible controller [0300]: nVidia Corporation GeForce 9500 GT [10de:0640] (rev a1)
     Subsystem: Micro-Star International Co., Ltd. Device [1462:1337]

[Bryce Harrington (bryce)]

Hi rupert-overall,

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.]

[ruru (ruru)]

I have not tinkered with the xorg.conf file (it is very boring) - attached here anyway...

[ruru (ruru)]

The attached files were generated after restarting (dots are present - as always). Attached screenshot shows how the dots are larger than a single pixel and are not restricted to desktop/panel/window.
They are also present on the login screen in the same pattern.

I hope all this info helps...

[Joseph Coffland (jcoffland)]

This bug also showed up recently on my Intrepid desktop install with nvidia-glx-180.

I would also like to note that if I mouse click on the dots they blink. This may be due to a screen redraw at the X layer. Moving a window over the dots will also make them blink. It seems that X firsts draws the correct content then a moment later the dots reappear, thereby causing the blinking.

[ejdo (erik-derohanian)]

same bug in animux, xfce, nvidia 173

[Paul van Genderen (paulvg)]

This also affects nvidia-185

[bugbot (bugbot)]

This bug report was filed against an old version of Ubuntu.
Can you confirm whether this is still an issue in natty?

If you don't mind, it would be very helpful if you could update the bug
report in launchpad to 'Fix Released' if it is no longer an issue for
you, or if it is still occurring under natty, please tag the bug 'natty'
so it's easier for us to track.

[Joseph Coffland (jcoffland)]

It is discouraging that the policy for many bugs is to simply wait until a newer version comes along so they can be downgraded and then closed with out any real investigation. I understand that developer resources are limited but shouldn't the bug remain 'Confirmed' until proven otherwise? This bug has been downgraded to 'Incomplete' with no evidence that the bug has been fixed other than a lack of recent reports.

I think the process should be to first ask if the bug still occurs. Then if there are no updates for some time or people can confirm that the bug no longer exists in the newer version then the status can be changed. Otherwise, it puts too much burden on the bug reporters to fight for a bug's status only to be repeatedly smacked down. Particularly the original reporter should have a chance to comment before the status is changed.

--- End rant ---

I haven't seen this bug in awhile because I stopped using the machine it occurred on. I believe it is related to faulty Nvidia hardware and/or drivers.

[Bryce Harrington (bryce)]

"I understand that developer resources are limited but shouldn't the bug remain 'Confirmed' until proven otherwise?"

Indeed, that is exactly what happened, the bug remained in confirmed state for a year and a half.

You have to understand that there are 400 bug reports against Ubuntu's X components, with around 200 bugs filed each month. It is very challenging to stay atop the new bugs, let alone the old ones. Fortunately, for natty we did in fact manage to stay more or less atop the bug flood this go around. You can see this from this graph:

  http://www.bryceharrington.org/Arsenal/Reports/ubuntu-x-swat/totals.svg

Truly investigating bugs takes a considerable amount of time. Deep analysis of a bug can take days of time, and you can see from the numbers (and like you mention in your post) we get way more requests from people to analyze their problems than we have manpower to do. So we focus effort on where we can provide the best bang for the buck. That usually translates to bugs that are confirmed as still issues with the current version (i.e. oneiric), that either we can reproduce ourselves or that have an active reporter who can do testing for us (and is willing to fight for the correct status for a bug), that are severe issues affecting a lot of people, and that we are likely going to be able to fix. (This latter point is why -nvidia and -fglrx bugs like this one tend to get less priority - because they're proprietary codebases, we fundamentally simply cannot legally or technically fix many problems, because the car's hood is welded closed to us.)

"I think the process should be to first ask if the bug still occurs. Then if there are no updates for some time or people can confirm that the bug no longer exists in the newer version then the status can be changed."

Well, basically what you describe is the process we're following here. Maybe it feels like a "smackdown" to have it marked 'Incomplete', but launchpad has special tracking features relating to Incomplete status, which is why I use it - this state (and only this state) has a flag to show when someone has replied to a bug after we've asked a question. Sorry you feel it to be a "smackdown", however as you can see, the process worked here - your "rant" got flagged as an answer and here I am. :-) Unfortunately Launchpad doesn't differentiate between answers from the original reporter vs. any random commenter (I _really_ wish it did...) but this process works.

Anyway, this is probably all TMI, but one rant deserves another I guess. ;-) I don't want you thinking we all just sit around on our thumbs all release and then spam everyone to retest, although I'm sure you must think that!

[Bryce Harrington (bryce)]

Regarding the bug itself... This does sound like either a driver problem or a problem with the hardware itself. If the latter, then the issue would show up in windows as well, which no one has mentioned so maybe we can assume it isn't.

From the spacing and pattern of the dots it looks like graphics memory writing failure. Like, the driver wasn't able to keep up with the data stream and skipped writing a pixel. This type of problem often will just "go away" after updating to a better performing software stack (either kernel or -nvidia or both). Thus, in this case having it re-tested against natty really is necessary as a first step.

If it does occur with the nvidia in natty, then this bug should be moved from 'nvidia-graphics-drivers-180' to 'nvidia-graphics-drivers' and tagged 'nvidia'. If you run "apport-collect 421048" on natty I think that should do the trick, but am not sure. From there, Canonical can flag the issue for NVIDIA to investigate.

[Joseph Coffland (jcoffland)]

Thank you for the concerned response. I have a good idea of what it's like to be on your end because I am on other projects. I don't mean to be a whiner.

I had a number of bugs related to an Nvidia card and multiple displays stemming back to Hardy. Most of the reports I filed or was involved with were eventually closed with out really being resolved as Ubuntu versions progressed. I guess in this case the main issue was probably a faulty Nvidia card/driver pair. The issues went away when I abandoned a particular machine with on board Nvidia. I should probably be yelling at Nivida for not owning up to and fixing the problems. I have an expensive laptop now paperweight due to these problems. I guess it's a bit of a sore spot for me.

Keep up the great work. Ubuntu is getting better and more reliable all the time.

[stop (whoopwhoop)]

I have this bug in 11.04

[Joseph Coffland (jcoffland)]

@woop What video card do you have? What driver version are you using? Your screenshot looks very similar to what I saw.

[stop (whoopwhoop)]

01:00.0 VGA compatible controller: nVidia Corporation GT200 [GeForce GTX 260] (rev a1) (prog-if 00 [VGA controller])
 Subsystem: ASUSTeK Computer Inc. Device 831a
 Flags: bus master, fast devsel, latency 0, IRQ 16
 Memory at f5000000 (32-bit, non-prefetchable) [size=16M]
 Memory at c0000000 (64-bit, prefetchable) [size=256M]
 Memory at f2000000 (64-bit, non-prefetchable) [size=32M]
 I/O ports at 8c00 [size=128]
 [virtual] Expansion ROM at f6480000 [disabled] [size=512K]
 Capabilities: [60] Power Management version 3
 Capabilities: [68] MSI: Enable- Count=1/1 Maskable- 64bit+
 Capabilities: [78] Express Endpoint, MSI 00
 Capabilities: [100] Virtual Channel
 Capabilities: [128] Power Budgeting <?>
 Capabilities: [600] Vendor Specific Information: ID=0001 Rev=1 Len=024 <?>

nvidia-current 270.41.06-0ubuntu1

[andornaut (andornaut)]

I have the same issue. I've reproduced it with two different cards based on the nvidia 210 chipset. I'll attach my lspci -vvnn output below.

Environment:

OS: Ubuntu 11.04
NVidia driver: 270.41.06

[lattenwald (qalexx)]

I have the same problem.

Ubuntu 11.04
NVidia driver 270.41.06

[lattenwald (qalexx)]

A screenshot (made with "import screenshot.png")

[lattenwald (qalexx)]

FYI: updated drivers from https://launchpad.net/~ubuntu-x-swat/+archive/x-updates, and the problem is gone for me.

[Jens Lehmann (jens-lehmann-t)]

I had exactly the same problem constantly since several years with different Ubuntu versions and different nvidia drivers.

After adding agp=off to the GRUB_CMDLINE_LINUX_DEFAULT in the grub config the strange black dots problem is completely gone! :-)

...maybe this helps someone too.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Onno Steenbergen (osteenbergen)]

For me it started to appear just last week. Running this version of NVidia drivers on a Dell Latitude 6520 using a 4200NVS on latest ubuntu and all updates:

nvidia-xconfig: version 313.09
(<email address hidden>) Tue Dec 4 22:54:46 PST
2012

Will try the agp=off

[Evan Bullock-Morales (evan-bm)]

I am having the same issue on 13.04 running on a Sony VAIO S Series core i7, Intel HD 4000 Graphics.

Periodically I get a screen full of small square dots scattered over the entire screen. They appear randomly, for about a
second, then are gone. Its not hardware; I dual boot windows 8, and I have no similar issue there.