USB-C ports dead in sys-usb after dom0 + template update on 2026-05-26

Hi all,

After this morning’s update, all USB-C ports on my laptop stopped working — devices get no power and aren’t detected. USB-A ports continue to work normally. Looking for help narrowing down whether this is a known regression or something specific to my setup.

Hardware

  • Laptop: AiStone X6AR57TY
  • BIOS: N.1.17PCS12
  • CPU: Intel Core Ultra 9 275HX (Arrow Lake-HX)
  • GPU: Intel Arrow Lake-S iGPU + NVIDIA GeForce RTX 5070 Ti Mobile (GB205M)
  • RAM: ~128 GB
  • Network: Intel Arrow Lake-S CNVi Wi-Fi, Realtek RTL8125 2.5GbE
  • Thunderbolt 4 controllers: 00:0d.0 (TB4 USB Controller, 8086:7ec0) and 00:0d.2 (TB4 NHI, 8086:7ec2)
  • PCH xHCI for USB-A: 80:14.0 (8086:7f6e)

Software

  • Qubes OS 4.3 (i3 as WM in dom0)
  • All three USB controllers attached to sys-usb
  • sys-usb template is on Fedora 43 (fc43)
  • Currently running dom0 kernel 6.19.14-1.qubes.fc41 (rollback to 6.19.5-1 did not help)

Symptom

  • USB-A ports: working
  • USB-C ports: no power, no enumeration. A USB-C device with a power LED stays off when plugged in.
  • Was working before this morning’s update. (Pre-existing issue with TB docks specifically, but plain USB-C devices were fine until today.)

What changed today (dnf history transaction 27)

Dom0:

  • Installed: kernel-1000:6.18.31-1.qubes.fc41 (not booted, still on 6.19.x)
  • Upgraded: qubes-core-admin-client 4.3.30 → 4.3.31, python3-qubesadmin 4.3.30 → 4.3.31
  • Upgraded: several firmware packages from 20260410 → 20260519 (intel-gpu-firmware, intel-audio-firmware, nvidia-gpu-firmware, amd-ucode-firmware, etc. — no thunderbolt-related blobs)

Inside sys-usb template (fc43):

  • qubes-core-agent-* 4.3.43-1
  • python3-qubesadmin 4.3.31-1
  • Firmware packages (20260519)

Investigation so far

PCI assignment is correct — qvm-pci shows both TB controllers attached to sys-usb with no-strict-reset=True.

Inside sys-usb:

  • lspci shows the TB4 USB controller (00:06.0) and NHI (00:07.0) present.
  • dmesg shows xhci_hcd 0000:00:06.0 initializing normally and a thunderbolt 0000:00:07.0: device links to tunneled native ports are missing! warning.
  • /sys/bus/thunderbolt/devices/ contains domain0 and 0-0 (host router).
  • Security level: user.
  • lsusb -t shows the TB USB controller buses present but with no devices ever appearing when something is plugged into a USB-C port.

Initially I thought it was a missing bolt daemon (was not installed in sys-usb post-update). Installed it; boltctl list returns empty even when a USB-C device is connected, suggesting the device never gets detected at the TB link layer in the first place.

Type-C subsystem state:

  • dom0 has /sys/class/typec/ with port0, port1, partners, plugs, cables.
  • sys-usb has ucsi_acpi, typec_ucsi, typec modules loaded but no /sys/class/typec/ directory.

So dom0 owns the Type-C ports (via UCSI/ACPI) and sys-usb owns the USB and TB controllers — which is presumably the working split. Something in today’s update appears to have broken the handoff between them.

Kernel command line (and why)

intel_pmc_core.disable=1 initcall_blacklist=intel_pmc_ssram_telemetry_driver_init module_blacklist=intel_pmc_ssram_telemetry rd.qubes.hide_all_usb

The intel_pmc_* parameters were added against kernel 6.19.4-1 (R4.3-rc4) to avoid a boot panic. They may no longer be required on current kernels, but they were already in place when USB-C was working yesterday, so I don’t think they’re related to today’s regression.

Question

Has anyone else seen USB-C break after a recent update on Qubes 4.3? Or is there a known interaction between qubes-core-admin-client 4.3.31 / qubes-core-agent 4.3.43 and Thunderbolt/USB-C passthrough that I should be aware of?

Happy to provide further logs (full dmesg, qvm-pci, etc.) if useful.

Thanks!

I did not have problems… however, based on the extensive info you pasted (well done, btw) - I would try to revert the recently upgraded firmware rpms in sys-usb!

I had the same issue but with USB-A ports instead, AND with 4.2. Just restarted and for now it seems OK. So, I guess it’s something with fedora and/or Linux 7.06x.

Ok, it solved itself. By updating the fedora 43 template today it installed kernel 7.0.8 (among some other packages) which seems to solve the problems on sys-usb.

1 Like