Weird graphical error on a i5 2500k integrated graphics

[hugonc...@gmail.com]

Hello,

First of all I'd like to thank all of you, both the development team and the community for the amazing work you've done. For a computer as old as mine, this is running great! I do have a problem though:

I'm running a i5 2500k, P8Z68-V LX, RX380 and 2 monitors, one 60hz another one 144hz. In order to have better performance, I'm using the integrated graphics to power the 60hz one and the GPU for the 144hz one. Everything works perfectly so far on the 144hz one, but the 60hz has had a problem since the beginning.

I'll try to describe it - there's a 45 degree "cut" on the screen, from the top left corner to the center, and then it continues to the right. Everything something changes around that "cut"'s area, there's a weird refresh, where at least one frame has things on top of it, etc. I tried looking up for ways to resolve it, the only thing I found was switching kernels, I tried using the unstable and others, everything works the same.

I've uploaded them to imgur so you can try and see the problem (the weird white lines usually disappear a few seconds after, or when I pull up the browser to that monitor)

(Forgive the quality, I had to use my phone in order to capture the problem)

http://imgur.com/a/6Yj8N

(take a look at the word "Displays" and you'll see the problem)

[Vít Šesták]

I am not sure if it is a GPU-related issue. It might be related to cable or the monitor itself. Few ideas to try:

* If they both are connected via the same cable type: Swap the two monitors. Or swap just the cables. Do you see the issue on the same monitor? Does the issue moved to the other monitor?
* Is the cable connected well on both sides?
* When using PrintScreen, do you see the artefact here?
* Do you see the same issue in, say, BIOS menu?

Regards,
Vít Šesták 'v6ak'

[Vít Šesták]

On June 17, 2017 10:19:40 PM GMT+02:00, Hugo Costa wrote:
>Sorry for taking so long to answer, I tested that today, my results
>were:
>
>1 - changing the cable had no effect. I used both another VGA cable as
>well
>as an HDMI cable, the result was the same. I swapped the cables and the
>issue moved to the other monitor, as expected
>2 - yes, I always did the testing with the cables tightned
>3 - I forgot this test, I'll try it again tomorrow
>4 - No, both on the BIOS and on other OSes (such as Linux Mint and
>Win10),
>it all works fine, I'm sure it's something to do with the integrated
>graphics compatibility with qubes :(
>
>Thanks for taking the time!
>

Since other OSes work well, it sounds like a driver issue. Since some other Linux distribution works well and this issue does not seem to be a common QubesOS issue (I don't remember any other occurrence of it on this mailinglist), I guess it is related to driver version. I suggest trying to update kernel to version from qubes-dom0-current-testing.

Regards,
Vít Šesták 'v6ak'