Bug 34220

Same here, the screen randomly blinks out in every 30/60/90th… seconds. (I also use Arch Linux)

My video card:
01:00.0 VGA compatible controller: nVidia Corporation NV17 [GeForce4 MX 440] (rev a3)

The error message on each blink from everything.log:
kernel: [drm] nouveau 0000:01:00.0: DDC responded, but no EDID for VGA-1
kernel: [drm] nouveau 0000:01:00.0: Load detected on output A

Package versions:
kernel26 2.6.37.3-1
libdrm 2.4.23-2
xf86-video-nouveau 0.0.16_git20101217-1
xorg-server 1.9.4-1
libgl 7.10.1-1
mesa 7.10.1-1

I reported this bug on Arch Linux bugtracker also:
https://bugs.archlinux.org/task/23213

*** Bug 34972 has been marked as a duplicate of this bug. ***

I have same problem, card is
01:00.0 VGA compatible controller: nVidia Corporation NV25 [GeForce4 Ti 4200] (rev a3)
OS is Gentoo.
I tried
Nouveau kernels 2.6.38-rc8-g7972f91 and 2.6.37-rc3-gd095e23,
xorg-servers 1.9.4, 1.9.5 and 1.10.0,
xf86-video-nouveau 0.0.16_pre20101130 and git 92db2bc.

Nothing helped, so I tried proprietary drivers 96.43.19 and noticed line
[    30.567] (WW) NVIDIA(GPU-0): Unable to read EDID for display device CRT-1
This lead me to trying i2cdetect and indeed on all 3 i2c adapters there is nothing on adress 0x50, where EDID should be. Any ideas?

Created attachment 44752 [details]
Xorg log with nouveau driver

Created attachment 44753 [details]
kernel log with nouveau driver

Created attachment 44754 [details]
Xorg log with nvidia driver

dmesg now outputs this:

[drm] nouveau 0000:01:00.0: Load detected on output A
[drm:drm_edid_block_valid] *ERROR* EDID checksum is invalid, remainder is 222
[drm] nouveau 0000:01:00.0: Load detected on output A
[drm:drm_edid_block_valid] *ERROR* EDID checksum is invalid, remainder is 222
[drm] nouveau 0000:01:00.0: Load detected on output A
[drm] nouveau 0000:01:00.0: Setting dpms mode 0 on vga encoder (output 0)
[drm] nouveau 0000:01:00.0: Setting dpms mode 0 on tmds encoder (output 2)

Turns out in my case problem was caused either by bad cable or dust in the monitor, after dusting monitor and manipulating with the cable, EDID is detected again, so sorry for the noise.

Anyway, should really Nouveau blink the monitor every 30s when it does not detect EDID?

I have the same problem with GeForce GTX 460 and nouveau. And indeed, even official nvidia driver doesn't find EDID for the screen which blinks with nouveau. Still, binary nvidia drives that monitor OK, so nouveau should too.

I think it has been fixed in upower with this commit:

 Add an option for polling dock devices in UPower.conf, defaulting to false
 http://cgit.freedesktop.org/upower/commit/?id=38199e44ff1e87b586d51d30b69d1f70e40d27f1

Sounds like the bug here is that nouveau causes the VGA display to blink on/off when checking whether it is connected. (And, to exacerbate the issue, there's something polling every 30s to see if the monitor is connected/what modes it has or something.)

Could someone with an affected monitor try to

cat /sys/class/drm/card0-VGA-0/status

(or whatever card/connector the thing is on)

and see if that action causes the monitor to blink on/off? If not, check /sys/.../modes and .../edid and .../dpms. I'm guessing at least one of those will cause the blinking issue.

My understanding of the VGA DDC stuff is that the monitor doesn't even have to be powered on in order to get at that stuff (I guess power is supplied in the VGA cable itself), much less have to try to show a picture. Although perhaps not all monitors comply with that ideal.

BTW, please be sure to test it out with the latest kernel, just in case this has been fixed.

(In reply to comment #11)
> Sounds like the bug here is that nouveau causes the VGA display to blink
> on/off when checking whether it is connected. (And, to exacerbate the issue,
> there's something polling every 30s to see if the monitor is connected/what
> modes it has or something.)
> 
> Could someone with an affected monitor try to
> 
> cat /sys/class/drm/card0-VGA-0/(In reply to comment #11)
> Sounds like the bug here is that nouveau causes the VGA display to blink
> on/off when checking whether it is connected. (And, to exacerbate the issue,
> there's something polling every 30s to see if the monitor is connected/what
> modes it has or something.)
> 
> Could someone with an affected monitor try to
> 
> cat /sys/class/drm/card0-VGA-0/status
> 
> (or whatever card/connector the thing is on)
> 
> and see if that action causes the monitor to blink on/off? If not, check
> /sys/.../modes and .../edid and .../dpms. I'm guessing at least one of those
> will cause the blinking issue.
> 
> My understanding of the VGA DDC stuff is that the monitor doesn't even have
> to be powered on in order to get at that stuff (I guess power is supplied in
> the VGA cable itself), much less have to try to show a picture. Although
> perhaps not all monitors comply with that ideal.
> 
> BTW, please be sure to test it out with the latest kernel, just in case this
> has been fixed.

I can confirm that blinking issue after cat /sys/class/drm/card0-VGA-0/status every ~30 seconds. I'm using nvidia 7300 gt. do you need more data?

(In reply to comment #12)
> I can confirm that blinking issue after cat
> /sys/class/drm/card0-VGA-0/status every ~30 seconds. I'm using nvidia 7300
> gt. do you need more data?

I'm very confused by your comment. Does cat'ing the status cause an immediate blink, yes or no? (Your comment, taken literally, seems to imply that things are all fine, then you cat the file, and then the blink happens every 30s. This seems unlikely.)

What kernel version are you using?

-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/xorg/driver/xf86-video-nouveau/issues/13.