[Bug 2168830] Re: Intel Cannon Point-LP xHCI controller dies on resume from deep suspend on Lenovo ThinkPad L490

Hi, Thanks for submitting a bug report. Does this happen every time or does it only happen occasionally? Was the occurrence on 2026-09-24 the first time this happened? -- You received this bug notification because you are subscribed to linux in Ubuntu. Matching subscriptions: Bgg, Bmail, Nb https://bugs.launchpad.net/bugs/2168830 Title: Intel Cannon Point-LP xHCI controller dies on resume from deep suspend on Lenovo ThinkPad L490 Status in linux package in Ubuntu: New Bug description: Kubuntu 26.04, kernel 7.0.0-34-generic (Ubuntu package 7.0.0-34.34) Machine: Lenovo ThinkPad L490, type 20Q5 BIOS: R0ZET60W (1.38) Controller: Intel Cannon Point-LP USB 3.1 xHCI [8086:9ded], PCI 0000:00:14.0 Sleep mode: deep (s2idle is also advertised) Lenovo USB-C dock attached (USB hubs 17ef:3072 and 17ef:3073). Repeated failure after suspend/resume, including automatic idle suspend. Latest occurrence: 2026-09-29, suspend 09:15:10 CEST, resume 09:27:10 CEST. Expected: USB devices reconnect and remain usable after resume. Actual: all USB devices disappear, including dock keyboard/mouse and internal Goodix fingerprint sensor. External monitor continues working. lsusb shows only USB2/USB3 root hubs. Internal keyboard and touchpad remain usable. Kernel messages: xhci_hcd 0000:00:14.0: xHCI host not responding to stop endpoint command xhci_hcd 0000:00:14.0: xHCI host controller not responding, assume dead xhci_hcd 0000:00:14.0: HC died; cleaning up usb 1-4.4.4: PM: dpm_run_callback(): usb_dev_resume returns -22 usb 1-4.4.4: PM: failed to resume async: error -22 Followed by disconnection of all USB devices. Recovery: unbind and rebind PCI 0000:00:14.0 from xhci_hcd restores USB without reboot. The issue also occurred on 2026-09-24. A separate fingerprint daemon hang occurs following USB re-enumeration; restarting fprintd fixes that symptom. A local sleep service now stops fprintd before suspend and starts it after resume, but the whole-controller failure still occurs. No controlled comparison with older kernels or suspend without the dock has been performed, so a kernel regression is not yet established. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2168830/+subscriptions