I previously asked about this on stackexchange, but that yielded no solution.
I originally noticed this on an Ubuntu VM, but now I can also reproduce this on my Debian VMs.
I am using two physical monitors, VMware Workstation Player 15, and a VM in full screen with a "Gnome unter X.org" session (A similar problem happens with Wayland, but that is fixed in recent versions).
When I use the "Cycle Multiple Monitors" feature a second screen appears as expected and initially works fine.
Clicking the button again however does not revert to a single screen configuration correctly.
Instead the second screen is completely black; mouse input does now longer work reliably, the pointer is offset somehow.
In case it matters: This happens with different hardware, (different physical machines, screens, cables) and different versions of VMware Workstation Player(15 and 17, all on Windows 10.
I am not currently able to test this outside of virtual machines.
Using pre-built packages I could discover that this was introduced between
6.1.0-20-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.85-1 (2024-04-11)
and
6.1.0-21-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.90-1 (2024-05-03).
I then used the linux-stable git source (which was surprisingly easy).
The problem appears in 6.1.88 but was good in 6.1.87 (built from source each)
I was able to git bisect the 6.1 series further to find the bad commit:
> first bad commit: [bcff1ed2ff1933be34b3737e5534c6650cd4e6e1] drm/vmwgfx: Fix crtc's atomic check conditional
The problem is still present in 6.1.106 (from Debian)
The problem also appears with a kernel build from tag: v6.11-rc6 and the official package for this version.
The problem also appears in the 5.15 series, but there I did not bisect further than Ubuntu pre-build packages.
Good: ubuntu-22-test 5.15.0-113-generic #123-Ubuntu SMP Mon Jun 10 08:16:17 UTC 2024
Bad: ubuntu-22-test 5.15.0-116-generic #126-Ubuntu SMP Mon Jul 1 10:14:24 UTC 2024
Considering that the commit mentions drm/vmwgfx which I understand to be (part of?) the video driver for virtual machines in vmware, I am quite confident that I found a cause.
What I do not know is: Is this an unintentional regression in the kernel? Or is the problem in user-space, maybe open-vm-tools-desktop? Or is x.org essentially no longer supported?
Should I fill a bug report? Where?
I originally noticed this on an Ubuntu VM, but now I can also reproduce this on my Debian VMs.
I am using two physical monitors, VMware Workstation Player 15, and a VM in full screen with a "Gnome unter X.org" session (A similar problem happens with Wayland, but that is fixed in recent versions).
When I use the "Cycle Multiple Monitors" feature a second screen appears as expected and initially works fine.
Clicking the button again however does not revert to a single screen configuration correctly.
Instead the second screen is completely black; mouse input does now longer work reliably, the pointer is offset somehow.
In case it matters: This happens with different hardware, (different physical machines, screens, cables) and different versions of VMware Workstation Player(15 and 17, all on Windows 10.
I am not currently able to test this outside of virtual machines.
Using pre-built packages I could discover that this was introduced between
6.1.0-20-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.85-1 (2024-04-11)
and
6.1.0-21-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.90-1 (2024-05-03).
I then used the linux-stable git source (which was surprisingly easy).
The problem appears in 6.1.88 but was good in 6.1.87 (built from source each)
I was able to git bisect the 6.1 series further to find the bad commit:
> first bad commit: [bcff1ed2ff1933be34b3737e5534c6650cd4e6e1] drm/vmwgfx: Fix crtc's atomic check conditional
The problem is still present in 6.1.106 (from Debian)
The problem also appears with a kernel build from tag: v6.11-rc6 and the official package for this version.
The problem also appears in the 5.15 series, but there I did not bisect further than Ubuntu pre-build packages.
Good: ubuntu-22-test 5.15.0-113-generic #123-Ubuntu SMP Mon Jun 10 08:16:17 UTC 2024
Bad: ubuntu-22-test 5.15.0-116-generic #126-Ubuntu SMP Mon Jul 1 10:14:24 UTC 2024
Considering that the commit mentions drm/vmwgfx which I understand to be (part of?) the video driver for virtual machines in vmware, I am quite confident that I found a cause.
What I do not know is: Is this an unintentional regression in the kernel? Or is the problem in user-space, maybe open-vm-tools-desktop? Or is x.org essentially no longer supported?
Should I fill a bug report? Where?
Statistics: Posted by FKuerten — 2024-10-09 09:27 — Replies 1 — Views 55