Dell wd25 docking station issues

[Sn0wSun]

Screen keeps blanking/switching having weird behavior

First looked how the kernel params were:
sudo rpm-ostree kargs
Added the following:
sudo rpm-ostree kargs --append='i915.enable_psr=0 i915.enable_fbc=0 iommu=pt pcie_ports=native'
Verified if they were added
sudo rpm-ostree kargs

Rebooted the system.

20250807_172611.mp4

Now the screen is stable

[Sn0wSun]

AI:

1️⃣ i915.enable_psr=0 — Disable Panel Self Refresh
What it is:
A power-saving feature in Intel’s iGPU driver where the panel refresh is paused if the image hasn’t changed, and the display fetches from internal memory instead of the GPU.

Why it breaks things:
On some Intel generations (including Meteor Lake) and with MST hubs like your Synaptics VMM8431 in the WD25 dock, PSR can cause link-training glitches or keep an external display “stuck” after mode changes.

Your case:
PSR was likely leaving the DP-7 link in a half-awake state → black screen.

2️⃣ i915.enable_fbc=0 — Disable Frame Buffer Compression
What it is:
Another Intel power-saving feature that compresses the framebuffer before sending it to the display.

Why it breaks things:
With MST (multi-stream transport) docks, compression can mess up bandwidth negotiation or cause timing issues, especially at higher resolutions like 2560×1440 over DP Alt-Mode.

Your case:
Disabling FBC ensures a clean, uncompressed stream → more reliable link.

3️⃣ iommu=pt — IOMMU pass-through mode
What it is:
Normally, the IOMMU (I/O Memory Management Unit) maps device DMA to system memory. pt (pass-through) mode tells it to skip that for most devices.

Why it helps here:
It removes a layer of address translation for PCIe devices (including your TB/USB4 dock’s GPU/DP tunnels). This can reduce latency and avoid quirks in some firmware/driver combos.

4️⃣ pcie_ports=native — Let Linux handle PCIe port services
What it is:
Determines who manages PCIe hotplug, error reporting, etc. — firmware (BIOS/ACPI) or the kernel.

Why it helps:
Some docks (especially Thunderbolt/USB4 with DP tunneling) hotplug their internal MST bridges at connect time. If the firmware isn’t managing ports correctly, the device won’t re-train the link; giving control to Linux fixes that.

Why all together fixed your issue
Your dock’s DisplayPort path is a DP Alt-Mode MST connection carried over USB4/TB.
Intel’s PSR/FBC were introducing timing quirks, and the IOMMU/PCIe settings ensured the MST hub and GPU could re-negotiate the link without the firmware getting in the way.
Once you killed those power-saving features and gave the kernel direct control of PCIe ports, the DP-7 link trained properly and stayed lit.

Trade-offs
Slightly higher power use on battery, because PSR/FBC save energy when idle.

iommu=pt means you lose some device isolation (only matters if you use VMs with PCIe passthrough or want maximum DMA protection).

Otherwise, perfectly fine to leave as-is — lots of Linux dock users run these permanently.

[L0g0ff]

@Sn0wSun what the status about this issue?

[Sn0wSun]

@Sn0wSun what the status about this issue?

No problems anymore after this has been set

[L0g0ff]

@Sn0wSun what the status about this issue?

No problems anymore after this has been set

I think it's quite tricky to make modifications on a docking station. Do you want a global fix for this? Or should I create a ujust file for these karg settings?

[Sn0wSun]

@Sn0wSun what the status about this issue?

No problems anymore after this has been set

I think it's quite tricky to make modifications on a docking station. Do you want a global fix for this? Or should I create a ujust file for these karg settings?

Ujust file, if there are problems you can use this to fix it