Confirmed across all available kernels and latest BIOS — not a version-
specific regression on this unit
System: ASUS ProArt P16 H7606WP, AMD Ryzen AI 9 HX 370, Radeon 890M
(Strix Point APU)
Did a full bisection across every kernel build available/installable on
this machine, plus a BIOS update. The exact same failure occurs in all
of them, with no variation:
amdgpu 0000:65:00.0: probe with driver amdgpu failed with error -22
Tested (all reproduce the identical failure):
linux-image-7.0.0-30-generic
linux-image-7.0.0-31-generic
linux-image-7.0.0-34-generic
linux-image-7.0.0-38-generic
linux-image-7.0.0-1013-oem
linux-image-7.0.0-1015-oem
Also updated BIOS from version 306 to the latest available, 307
(2026/09/23) — no change, identical error.
Additional detail that may help narrow this down — lspci -nnk -s 65:00.0
-vv shows the GPU's PCI class as [0380] ("Display controller - Other")
rather than the usual [0300] ("VGA compatible controller"):
65:00.0 Display controller [0380]: Advanced Micro Devices, Inc. [AMD/ATI] Strix
[Radeon 880M / 890M] [1002:150e] (rev c1)
Subsystem: ASUSTeK Computer Inc. Device [1043:12d4]
Region 0: Memory at 8a10000000 (64-bit, prefetchable) [size=256M]
Region 2: Memory at dc200000 (64-bit, prefetchable) [size=2M]
Region 4: I/O ports at e000 [disabled] [size=256]
Region 5: Memory at dc700000 (32-bit, non-prefetchable) [size=512K]
Kernel modules: amdgpu
Boot log also shows ATPX hybrid-graphics detection running successfully
just before the probe failure:
amdgpu: vga_switcheroo: detected switching method \_SB_.PCI0.GPPA.VGA_.ATPX
handle
amdgpu: ATPX version 1, functions 0x00000001
amdgpu: ATPX Hybrid Graphics
amdgpu: Virtual CRAT table created for CPU
amdgpu: Topology: Add CPU node
amdgpu 0000:65:00.0: probe with driver amdgpu failed with error -22
No PCI resource/BAR allocation errors anywhere in dmesg — the device
enumerates cleanly, so this doesn't look like a BIOS resource-allocation
problem. The failure happens very early in amdgpu's probe, before any
firmware loading is logged.
Given the failure is 100% reproducible across every kernel build and
BIOS version available to me, this looks less like a regression between
kernel versions and more like the amdgpu driver not yet correctly
handling this specific device/subsystem ID combination (1002:150e,
subsystem 1043:12d4) in its hybrid-graphics (ATPX) path. As a
consequence, no display output works at all on this machine under Linux
— including external monitors via the physical ports, since the display
engine never initializes.
Happy to test any debug patches, kernel command-line options, or provide
further logs (full dmesg, amdgpu_dc= debug output, etc.) if that would
help diagnose this further.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2135274
Title:
amdgpu driver probe fails with error -22 on kernel 6.17.0-8 (Strix
Point / Radeon 890M)
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2135274/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
