Public bug reported: Hardware: Lenovo ThinkPad T14s Gen 6, machine type 21N10000US, Snapdragon X Elite X1E-78-100, BIOS N42ET93W (2.23). Wi-Fi and Bluetooth card WCN7850, PCI 17cb:1107, subsystem 17aa:e0e9.
Kernel: linux-image-7.3.0-10-generic 7.3.0-10.10 (stonking-updates), Secure Boot on. Booted with Ubuntu's own signed kernel packages from a live image built from Ubuntu 26.04 arm64 packages. What happens: there is no Bluetooth controller. btmgmt info shows "Index list with 0 items". The kernel log repeats: Bluetooth: hci0: command 0xfc00 tx timeout Bluetooth: hci0: Reading QCA version information failed (-110) Bluetooth: hci0: Retry BT power ON:1 Bluetooth: hci0: AOSP get vendor capabilities (-110) hci_uart_qca also prints "supply vddio not found, using dummy regulator" for all six supplies. gpio116 (W_DISABLE2#) reads "out low", and /sys/kernel/debug/pwrseq shows connector-1's uart-enable unit with a count of 0. The same machine and image with 26.04's linux-image-7.0.0-42-generic works: Bluetooth: hci0: setting up wcn7850 Bluetooth: hci0: QCA Downloading qca/hmtbtfw20.tlv Bluetooth: hci0: QCA Downloading qca/hmtnv20.b112 Bluetooth: hci0: QCA setup on UART is completed The controller comes up and a scan finds devices. Cause: 7.3.0-6 (LP: #2167217, "26.10 Snapdragon X Elite: Sync concept kernel changes") replaced the T14s device tree's WCN7850 power unit description with the SAUCE "Describe PCIe M.2 1620 LGA connector" (connector-1, pcie-m2-1620-lga-connector / pcie-m2-e-connector, W_DISABLE1# on gpio117, W_DISABLE2# on gpio116, port 0 to pcie4_port0, port 3 to uart14). pwrseq-pcie-m2 then creates the bluetooth node (qcom,wcn7850-bt) and its serdev, and holds W_DISABLE2# asserted until a consumer enables its "uart" target. hci_qca in 7.3 asks for the "bluetooth" target on the serdev device itself. pwrseq-pcie-m2 matches only the connector's remote port parents (uart14 and pcie4_port0), not the bluetooth child, so the lookup returns -EPROBE_DEFER. hci_qca takes that as BT_EN pulled up in hardware, uses dummy regulators and never powers the chip. Fix: the hci_qca half of the same M.2 work, which asks for the "uart" target on the controller when the device has an OF graph: [v7,1/8] Bluetooth: hci_qca: Add M.2 Bluetooth device support using pwrseq (Loic Poulain, 2026-10-05) https://patchwork.kernel.org/project/bluetooth/patch/[email protected]/ series: https://patchwork.kernel.org/series/1179616/ Upstream also has a different approach under review: [1/5] Bluetooth: hci_qca: Add auxiliary driver for PCIe M.2 modules (Manivannan Sadhasivam, 2026-09-15) https://patchwork.kernel.org/project/bluetooth/patch/[email protected]/ Neither is in mainline or bluetooth-next yet. Since stonking carries the connector device tree as SAUCE, please carry the matching hci_qca patch with it, or keep the WCN7850 power unit description for the T14s until upstream settles. Other machines: in 7.3.0-10's built-in device trees, the ASUS Zenbook A16 (UX3607OA) uses the same E-key connector description, and the X1E80100 CRD no longer has a bluetooth node. Neither was tested here. ** Affects: linux (Ubuntu) Importance: Undecided Status: New ** Tags: arm64 stonking x1e -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2172779 Title: [x1e80100] ThinkPad T14s Gen 6: no Bluetooth on 7.3 (M.2 connector device tree without the hci_qca pwrseq patch) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2172779/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
