[Bug 221552] New: btmtk: MT7921 (USB 0e8d:e020 / PCI 14c3:7920) Bluetooth broken since 7.0.7 - fix for 0489:e0e2 does not cover this PID

[bugzilla-daemon@kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=221552

            Bug ID: 221552
           Summary: btmtk: MT7921 (USB 0e8d:e020 / PCI 14c3:7920)
                    Bluetooth broken since 7.0.7 - fix for 0489:e0e2 does
                    not cover this PID
           Product: Drivers
           Version: 2.5
          Hardware: All
                OS: Linux
            Status: NEW
          Severity: blocking
          Priority: P3
         Component: Bluetooth
          Assignee: linux-bluetooth@vger.kernel.org
          Reporter: gagnieux.virgil@proton.me
        Regression: No

Created attachment 310159
  --> https://bugzilla.kernel.org/attachment.cgi?id=310159&action=edit
dmesg kernel (from 7.0.9-hardened1-1-hardened)

Regression: Bluetooth controller fails to initialise on a MediaTek MT7921
(Legion Pro 5 16ADR10, Realtek-branded module) since kernel 7.0.7.
Last known good: 7.0.6. Same dmesg symptom as the regression already
patched for USB ID 0489:e0e2, but the fix does NOT resolve it for this
device (USB 0e8d:e020), which suggests the patch missed a code path or
device-ID entry.

Hardware
--------
lspci -nnk:
  03:00.0 Network controller [0280]: MEDIATEK Corp. Device [14c3:7920]
        DeviceName: Realtek
        Subsystem: Lenovo Device [17aa:e020]
        Kernel driver in use: mt7921e

lsusb:
  Bus 003 Device 006: ID 0e8d:e020 MediaTek Inc. Wireless_Device

Kernels tested
--------------
7.0.6                    : Bluetooth works
7.0.8 / 7.0.9 (mainline) : broken
7.0.9-hardened1-1        : broken (linux-hardened — current)
  This build includes the backport
  https://github.com/anthraxx/linux-hardened/commit/<…81d80fcd09>
  which adds the fix for USB ID 0489:e0e2. It does NOT fix 0e8d:e020.

uname -r : 7.0.9-hardened1-1-hardened

dmesg (from kernel 7.0.9, relevant lines)
-----------------------------------------
[   23.898011] Bluetooth: Core ver 2.22
[   23.898032] NET: Registered PF_BLUETOOTH protocol family
[   23.898033] Bluetooth: HCI device and connection manager initialized
[   23.898041] Bluetooth: HCI socket layer initialized
[   23.898043] Bluetooth: L2CAP socket layer initialized
[   23.898045] Bluetooth: SCO socket layer initialized
[   24.800153] Bluetooth: hci0: HW/SW Version: 0x008a008a, Build Time:
20260224111231
[   24.922472] Bluetooth: hci0: Failed to send wmt func ctrl (-22)
[   24.922477] Bluetooth: hci0: HCI Enhanced Setup Synchronous Connection
                command is advertised, but not supported.
[   25.517793] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
[   25.517796] Bluetooth: BNEP filters: protocol multicast
[   25.517800] Bluetooth: BNEP socket layer initialized

Userspace state
---------------
bluetoothctl list  → (no output)
bluetoothctl show  → No default controller available

Module
------
modinfo btmtk:
  filename:
/lib/modules/7.0.9-hardened1-1-hardened/kernel/drivers/bluetooth/btmtk.ko.zst
  firmware: mediatek/mt7925/BT_RAM_CODE_MT7925_1_1_hdr.bin
  firmware: mediatek/BT_RAM_CODE_MT7961_1_2_hdr.bin
  firmware: mediatek/BT_RAM_CODE_MT7922_1_1_hdr.bin
  firmware: mediatek/mt7668pr2h.bin

Versions (Arch Linux)
---------------------
linux-firmware           : 20260410-1
linux-firmware-mediatek  : 20260410-1
bluez                    : 5.86-6

Steps to reproduce
------------------
1. Boot kernel >= 7.0.7 on a system with MediaTek 14c3:7920 / USB 0e8d:e020.
2. Bluetooth controller fails to initialise; -EINVAL (-22) on wmt func ctrl.

Expected
--------
Controller comes up, same as on 7.0.6.

Related
-------
- Arch forum thread: https://bbs.archlinux.org/viewtopic.php?id=313552
- Same symptom, different PID 0489:e0d8:
  https://bbs.archlinux.org/viewtopic.php?id=313561
- Existing fix that targets 0489:e0e2 only:
 
https://github.com/anthraxx/linux-hardened/commit/d019930b0049fc2648a6b279893d8ad330596e81
  (does not cover USB ID 0e8d:e020 — a second device-ID / code-path entry
  appears to be needed)

I can test patches.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are the assignee for the bug.