You are not logged in.
As the title says, my screen goes black for a couple of second while playing video games or going to specific websites. It's also random from what I see (sometimes it can happen every 5 minutes or every second). I have observed this in ZDoom, MInecraft, World of Tanks, War Thunder, Half Life, Half Life 2; still haven't observed this in CS:GO. My PC specs are:
CPU: Ryzen 5 5600
GPU: Radeon RX 560
RAM: 16GB DDR4
Offline
For the usual suspects, add
amdgpu.aspm=0 amdgpu.runpm=0 amdgpu.bapm=0 amdgpu.audio=0 pcie_aspm=off to the kernel commandline
Also elaboarate on the environment (display server) and test the behavior on a (in doubt: much) lower resolution.
Online
I added the args and I guess it happens less, but that could only be placibo effect. For the server, xorg and herbstluftwm.
Offline
xrandr -q and please post your xorg log, https://wiki.archlinux.org/title/Xorg#General
Did you try the behavior w/ lower resolutions?
Online
I attempted several lower resolutions, but it didn't change much. Is there a way to upload files or post text in a more compact form here since the two logs are quite long?
Offline
Also 1st link below
Online
https://ufile.io/f/gqs0p
Here's the both logs (1 and 2)
Offline
So on request I linked you two non-annoying pastebin services and you figured you use one that requires multiple clicks, has time and speed restrictions, wants me to register and forces you to download the file…
What's next? Capchas?
The ouput gets polled a lot
[ 73.653]
[ 84.818] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 110.192] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 145.672] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 437.191] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 437.200] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 437.209] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 437.954] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 437.962] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 437.971] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 437.980] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 455.616] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 455.625] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14650.293] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14744.040] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14744.047] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14750.343] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14764.737] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14792.960] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14792.970] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14792.979] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14804.406] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14804.415] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14804.423] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14805.270] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14805.276] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14805.285] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14805.294] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14817.224] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14817.232] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14827.876] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14827.884] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14827.890] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14842.588] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14842.593] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14842.867] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14842.877] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14842.886] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14842.892] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14854.512] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14854.518] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14876.098] (II) modeset(0): EDID vendor "BNQ", prod id 30787
[ 14884.027] (II) modeset(0): EDID vendor "BNQ", prod id 30787Assuming that's not you running "xrandr -q" all the time, do the timestamps match w/ the blackouts?
(You can also tail the file and se whether it gets a new entry w/ each blackout)
The also come in bursts (up to 6 times a second) - do you run those games in fullscreen mode? Can you run them windowed? Does that also cause the blackouts?
Are zdoom & minecraft the native games or a windows version through steam?
Online
This happens in both windowed and full screen mode; although I discovered something, if I run it inside of a frame and expand that frame to as large as it can be (not the whole screen sadly) it doesn't happen. ZDoom and Minecraft are run natively, but WoT and WT are run through Steam Proton.
Offline
It is also possible that I didn't add the args correctly or to the correct place. So where am I supposed to add them and how?
Offline
They don't apply to the Xorg.0.log you posted, https://wiki.archlinux.org/title/Kernel_parameters
Online
I added them to GRUB_CMDLINE_LINUX_DEFAULT="" and did grub-mkconfig, but doesn't seem to have changed anything. Should I try other options?
Offline
You can see whether the apply w/
cat /proc/cmdlineDo you use grub to boot at all?
Did you forget to mount the /boot partition?
sudo -v # to not have to enter the pw below
(cat /proc/cmdline; lsblk -f; sudo cat /boot/grub/grub.cfg) | curl -F 'file=@-' 0x0.stOnline
Here is what cat command returns
"BOOT_IMAGE=/boot/vmlinuz-linux-zen root=UUID=c1947de3-0ce1-4021-be4f-ad618c0f2bef rw amdgpu.aspm=0 amdgpu.runpm=0 amdgpu.bapm=0 amdgpu.audio=0 pcie_aspm=off"
Offline
This shows that the parameters now apply.
Unfortunately the PM settings don't prevent the blackouts.
Do you have a spare monitor?
Also
This happens in both windowed and full screen mode … if I run it inside of a frame and expand that frame to as large as it can be … it doesn't happen
can you explain how the maximum size frame differs from using the windowed mode?
Do you use a compositor (picom) next to herbstluftwm?
I don't really see how either the WM or DPMS would explain the output polling frequency in the log, but nevertheless, try the behavior w/ openbox and
xset -dpms # + turns dpms back onOnline
I do not have a spare monitor, but one thing that might be an issue is that I use a VGA to DVI adapter to connect my monitor to the GPU.
When I run it in windowed mode it covers everything aside from the taskbar (tint2), but when i create another frame, the window only covers around 80% of the screen with the other 20% being given to the other frame. I do not use a picom.
I will test the last part now.
Offline
It's gonna be the adapter.
Post the output of"xrandr --verbose", maybe we can push down the signal.
Online
Here's the output of "xrandr --verbose" (https://pastebin.com/wNJXdm2i), also I managed to get a spare monitor from my neighbor and so far I haven't seen the black screen so far. I will test it more.
Offline
1440x900 (0x4a) 106.500MHz -HSync +VSync *current +preferred
h: width 1440 start 1520 end 1672 total 1904 skew 0 clock 55.93KHz
v: height 900 start 903 end 909 total 934 clock 59.89HzThat's the naive modeline, so you could try
◉ cvt12 1440 900 60 -r
# 1440x900 @ 60.000 Hz Reduced Blank (CVT) field rate 59.901 Hz; hsync: 55.469 kHz; pclk: 88.75 MHz
Modeline "1440x900_60.00_rb1" 88.75 1440 1488 1520 1600 900 903 909 926 +hsync -vsyncor
◉ cvt12 1440 900 60 -b
# 1440x900 @ 60.000 Hz Reduced Blank (CVT) field rate 60.000 Hz; hsync: 55.560 kHz; pclk: 84.45 MHz
Modeline "1440x900_60.00_rb2" 84.45 1440 1448 1480 1520 900 912 920 926 +hsync -vsynchttps://wiki.archlinux.org/title/Xrandr … esolutions which will reduce the signal by 20% for the rb2 modeline
The more you reduce the blanking interval, the more likely you'll experience tearing (despite vsync), but I'm not sure whether that's a concern anyway (b/c or the adapter you might already not get a reliable vblank signal from the monitor)
Online
Just tried both and tested for a bit, same issue and alongside that they were horribly stretched.
Offline
Those are supposed to be 16:10 1440x900 modelines, like the one active in the xrandr output you posted (attached was a Samsung Syncmaster from 2007)
How horribly exactly are they "stretched" and in which direction?
Online
In the way that like 20-30% of the screen goes beyond the edges.
Offline
Did you check the monitor OSD?
It might have interpreted the lower signal as 1024x768… though if it still drops out, that ultimately doesn't matter.
What's the exact monitor model?
And did your neighbours monitor hold up, btw?
Online
I've tested the monitor on multiple games and it seems that the issue persists. Could it be a driver issue?
Offline
Do you mean to say that you get the same behavior w/ your neighbours monitor (and hopefully no adapter)?
Online