@Danielduel If you find some time, could you test this?
Hello @bblacher, at a glance this reads like a video driver or graphics card hardware limitation. It may be worthwhile to also mention this issue to your video driver vendor if it hasn't been already.
@mSparks43 Yes, I have this issue since 1.25 and it sounds nothing like my problem.
@kisak-valve Do you think the amd/drm Gitlab is the place to go for mentioning this?
@mSparks43 Yes, I have this issue since 1.25 and it sounds nothing like my problem.
Have you checked which monitor vrcompositor is selecting when it fails, because if it tries to select a monitor that isn't disabled it fails since 1.25.
No I have not, I don't think it's selecting any monitor, I just think that the is some bug preventing it from entering direct mode.
Also, I don't have this issue since 1.25, I've always had this issue. It's just that 1.25 was the first version I tried. @mSparks43
No I have not, I don't think it's selecting any monitor, I just think that the is some bug preventing it from entering direct mode.
Also, I don't have this issue since 1.25, I've always had this issue. It's just that 1.25 was the first version I tried. @mSparks43
#609 is when I switch my third monitor on it selects the wrong monitor and fails (unless I disable the monitor it is trying to select, then it succeeds in using it for the VR output instead of the headset)
Succeeds (in using the Dell 24" as a direct mode VR headset)
Fails to start VR on any display.
@mSparks43 For me it doesn't matter which monitor I disable, I just have to disable one of them.
@mSparks43 For me it doesn't matter which monitor I disable, I just have to disable one of them.
does that disable it, or physically switch it off? different things happen for me depending on which I do.
I tracked "failing with three screens active" down to
.local/share/Steam/logs/vrcompositor.txt
with the screen disabled it uses it instead of the headset
with the screen off it uses the headset and works fine
with the screen on, it fails because instead of
[Info] - Found candidate direct display as RandR output 0x1ca
when it uses the headset and works fine
its picking an active display
[Info] - Found candidate direct display as RandR output 0x1bd
and cannot acquire an xlib display that is in use.
So I guess the crux of it being a duplicate or not, is whether it is actually trying to select the headset when it fails (is it picking the same display when it works and when it does not work).
Latest SteamVR, updated Arch
HDMI 1080p60
HDMI 1080p60
DP (1.4?) ultrawide 1440p120 (5120x1440)
Another DP - Index HMD
The problem is not present
@bblacher
@Danielduel
I do still have it, but as I said, my main screen is 2160p, so that may make the difference. Thank you for testing though :)
Nothing extracted yet.
Describe the bug
My setup is as follows:
7900xtx
DP-1 3840x2160@144Hz
DP-2 1920x1080@144Hz
HDMI-A-1 1920x1080@60Hz
If I try to start SteamVR it fails with error -203 and 307.

In order to use SteamVR in direct mode I have to use either
swaymsg output DP-2 power offorswaymsg output HDMI-A-1 power off.To Reproduce
Run a similar screen combination and see it fail.
Expected behavior
SteamVR should start normally
System Information
Additional context
journal of me trying to start SteamVR twice with all three screens connected. If I use
swaymsg output xx power offto disable at least one screen SteamVR starts successfully.This is no duplicate of https://github.com/ValveSoftware/SteamVR-for-Linux/issues/552, I tried turning down the refresh rate.