AMD eGPU detected but not starting

Hi Everyone,

I have a Lenovo Legion BoostStation TB3 eGPU enclosure. When new it came with an RTX2060, this worked fine, showed up and I got a video output.

I have recently replaced this RTX2060 with an RX5600XT for better Linux support. When I received the 5600 I first put it inside my main PC to confirm it al worked fine, it did. I then put it in the eGPU enclosure. However after doing that I cannot for the life of me get it to start and work.

The TB dock is connected and authorised, the GPU shows up in lspci and the amdgpu driver is listed but not loaded:

07:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Navi 10 [Radeon RX 5600 OEM/5600 XT / 5700/5700 XT] (rev ca)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Navi 10 [Radeon RX 5600 OEM/5600 XT / 5700/5700 XT]
        Kernel modules: amdgpu

I did a quick bit of digging and a search for amdgpu in dmesg yielded this:

[    9.255922] [drm] amdgpu kernel modesetting enabled.
[    9.256040] amdgpu: CRAT table disabled by module option
[    9.256042] amdgpu: Virtual CRAT table created for CPU
[    9.256048] amdgpu: Topology: Add CPU node
[    9.256150] amdgpu 0000:07:00.0: enabling device (0000 -> 0003)
[    9.759720] amdgpu 0000:07:00.0: amdgpu: get invalid ip discovery binary signature
[    9.759723] [drm:amdgpu_discovery_set_ip_blocks [amdgpu]] *ERROR* amdgpu_discovery_init failed
[    9.760018] amdgpu 0000:07:00.0: amdgpu: Fatal error during GPU init
[    9.760020] amdgpu 0000:07:00.0: amdgpu: amdgpu: finishing device.
[    9.760775] amdgpu: probe of 0000:07:00.0 failed with error -22

I did some researching on various elements of this output but was unable to find anything that seems to relate to my exact issue.I considered that maybe the GPU was unsupported in some wat however Lenovo sold a version of this enclosure that came with a 5700XT so I would be surprised if it was incompatble with the 5600XT. Has anyone got any ideas?

Hardware:
2022 LG Gram 16
i7-1260p
16GB RAM
Lenovo Legion BoostStation Thunderbolt 3 GPU enclosure
MSI RX5600XT Gaming X

Software:
Fully up to date endeavourOS, Kernel 6.5.5-arch1-1

Thanks in advance,
eob

EDIT: A few more dmesg outputs:

[    3.336257] ACPI: bus type thunderbolt registered
[    6.452300] thunderbolt 0-0:1.1: new retimer found, vendor=0x8087 device=0x15ee
[    7.116871] thunderbolt 0-1: new device found, vendor=0x108 device=0x1907
[    7.116875] thunderbolt 0-1: Lenovo Legion BoostStation
[    7.663319] thunderbolt 0-301: new device found, vendor=0x108 device=0x1908
[    7.663324] thunderbolt 0-301: Lenovo Legion BoostStation
[    8.081166] pci 0000:07:00.0: [1002:731f] type 00 class 0x030000
[    8.081200] pci 0000:07:00.0: reg 0x10: [mem 0x00000000-0x0fffffff 64bit pref]
[    8.081221] pci 0000:07:00.0: reg 0x18: [mem 0x00000000-0x001fffff 64bit pref]
[    8.081233] pci 0000:07:00.0: reg 0x20: [io  0x0000-0x00ff]
[    8.081245] pci 0000:07:00.0: reg 0x24: [mem 0x00000000-0x0007ffff]
[    8.081257] pci 0000:07:00.0: reg 0x30: [mem 0x00000000-0x0001ffff pref]
[    8.081498] pci 0000:07:00.0: PME# supported from D1 D2 D3hot D3cold
[    8.081737] pci 0000:07:00.0: 8.000 Gb/s available PCIe bandwidth, limited by 2.5 GT/s PCIe x4 link at 0000:00:07.0 (capable of 252.048 Gb/s with 16.0 GT/s PCIe x16 link)
[    8.081879] pci 0000:07:00.0: Adding to iommu group 18
[    8.081889] pci 0000:07:00.0: vgaarb: bridge control possible
[    8.081890] pci 0000:07:00.0: vgaarb: VGA device added: decodes=io+mem,owns=none,locks=none
[    8.081932] pci 0000:07:00.1: [1002:ab38] type 00 class 0x040300
[    8.081957] pci 0000:07:00.1: reg 0x10: [mem 0x00000000-0x00003fff]
[    8.082166] pci 0000:07:00.1: PME# supported from D1 D2 D3hot D3cold
[    8.082334] pci 0000:07:00.1: Adding to iommu group 18

On the main pc you have only intel with internal graphics?

Edit: It looks like it’s loading with amdgpu and failing to initialize. What exactly is happening ? Doesn’t boot up to the screen?

https://wiki.archlinux.org/title/External_GPU

The main PC I first tested the card in is an AMD system with no integrated graphics. The laptop I am trying to use the eGPU on just has intel integrated.

The eGPU powers up and fans spin. the ethernet and usb ports in the dock works fine etc, however the GPU does nothing, no video out or display detection. The logs I posted are basically the only evidence of the card actually existing.

I took a look at that arch egpu page but nothing works, DRI_PRIME always just uses the onboard intel regardless of value

When you boot it up you are still able to get into a tty? What does it show for inxi -Ga