[Bug 220419] New: Bluetooth stops working on Intel BE201 a few seconds after connecting a BT peripheral

[Bug 220419] New: Bluetooth stops working on Intel BE201 a few seconds after connecting a BT peripheral

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=220419

            Bug ID: 220419
           Summary: Bluetooth stops working on Intel BE201 a few seconds
                    after connecting a BT peripheral
           Product: Drivers
           Version: 2.5
    Kernel Version: 6.15
          Hardware: Intel
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P3
         Component: Bluetooth
          Assignee: linux-bluetooth@vger.kernel.org
          Reporter: anton@ekblad.cc
        Regression: Yes
           Bisected 6ed83047389c63a8d7712b0bf19eb38a6e4dc68e
         commit-id:

Since 6.15, Bluetooth stops working a few dozen seconds after connecting one or
more BT peripherals on Intel BE201.

"Stops working" means that all BT connections time out, Blueman and anything
else that interacts with Bluetooth hangs, and a series of messages like this
end up in dmesg:

Aug 05 21:01:40 albedo kernel: Bluetooth: hci0: Failed to send frame (-62)
Aug 05 21:01:40 albedo kernel: Bluetooth: hci0: sending frame failed (-62)
Aug 05 21:01:41 albedo kernel: Bluetooth: hci0: Failed to send frame (-62)
Aug 05 21:01:41 albedo kernel: Bluetooth: hci0: sending frame failed (-62)
Aug 05 21:01:43 albedo kernel: Bluetooth: hci0: command tx timeout
Aug 05 21:02:10 albedo kernel: Bluetooth: hci0: Failed to send frame (-62)
Aug 05 21:02:10 albedo kernel: Bluetooth: hci0: sending frame failed (-62)
Aug 05 21:02:40 albedo kernel: Bluetooth: hci0: Failed to send frame (-62)
Aug 05 21:02:40 albedo kernel: Bluetooth: hci0: sending frame failed (-62)
Aug 05 21:03:10 albedo kernel: Bluetooth: hci0: Failed to send frame (-62)
Aug 05 21:03:10 albedo kernel: Bluetooth: hci0: sending frame failed (-62)
Aug 05 21:03:40 albedo kernel: Bluetooth: hci0: Failed to send frame (-62)
Aug 05 21:03:40 albedo kernel: Bluetooth: hci0: sending frame failed (-62)
Aug 05 21:04:40 albedo kernel: Bluetooth: hci0: link tx timeout
Aug 05 21:04:40 albedo kernel: Bluetooth: hci0: killing stalled connection
dc:2c:26:3e:58:cf
Aug 05 21:04:40 albedo kernel: Bluetooth: hci0: Failed to send frame (-62)
Aug 05 21:04:40 albedo kernel: Bluetooth: hci0: sending frame failed (-62)
Aug 05 21:04:42 albedo kernel: Bluetooth: hci0: command tx timeout
Aug 05 21:04:43 albedo kernel: Bluetooth: hci0: Failed to send frame (-62)
Aug 05 21:04:43 albedo kernel: Bluetooth: hci0: sending frame failed (-62)
Aug 05 21:04:45 albedo kernel: Bluetooth: hci0: command tx timeout

Some other observations that might be helpful:
- Problem still exists in 6.16 and 6.15.9.
- WiFi is not affected.
- Unloading and reloading btintel_pcie and btintel "fixes" the problem for
another few dozen seconds.
- Turning off power management for BE201 doesn't seem to affect the problem.
- Actively streaming audio over Bluetooth seems to delay the problem.
- Tested with all possible permutations of Sony WH-1000XM6 headphones, Logitech
MX Master 3S mouse, and Keychron K7 keyboard.
- Bisecting shows the problem was introduced by commit
6ed83047389c63a8d7712b0bf19eb38a6e4dc68e.
- Hardware is a Dell XPS 13 9350 (2024, Lunar Lake) with Intel BE201 WiFi/BT.
- Firmware version is 20250708.

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

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

[Bug 220419] Bluetooth stops working on Intel BE201 a few seconds after connecting a BT peripheral

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=220419

Luiz Von Dentz (luiz.dentz@gmail.com) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |luiz.dentz@gmail.com

--- Comment #1 from Luiz Von Dentz (luiz.dentz@gmail.com) ---
As far I know BE201 still use btusb, not really sure why a change to
btintel_pcie would have any effect on a device over btusb but perhaps there is
something changed that affect btintel which is shared between both.

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

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

[Bug 220419] Bluetooth stops working on Intel BE201 a few seconds after connecting a BT peripheral

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=220419

--- Comment #2 from Paul Menzel (pmenzel+bugzilla.kernel.org@molgen.mpg.de) ---
Details for the bisected commit [1]:

$ git describe 6ed83047389c63a8d7712b0bf19eb38a6e4dc68e
v6.14-rc7-1458-g6ed83047389c
$ git show 6ed83047389c63a8d7712b0bf19eb38a6e4dc68e
commit 6ed83047389c63a8d7712b0bf19eb38a6e4dc68e
Author: Kiran K <kiran.k@intel.com>
Date:   Tue Feb 25 17:13:10 2025 +0530

    Bluetooth: btintel_pcie: Setup buffers for firmware traces

    This patch allocates the host memory which is used by controller to dump
    the firmware traces. The memory needs to be shared with controller via
    context information.

    Co-developed-by: Vijay Satija <vijay.satija@intel.com>
    Signed-off-by: Vijay Satija <vijay.satija@intel.com>
    Signed-off-by: Kiran K <kiran.k@intel.com>
    Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
    Signed-off-by: Arnd Bergmann <arnd@arndb.de>
[…]

https://web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=6ed83047389c63a8d7712b0bf19eb38a6e4dc68e

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

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

[Bug 220419] Bluetooth stops working on Intel BE201 a few seconds after connecting a BT peripheral

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=220419

--- Comment #3 from Anton Ekblad (anton@ekblad.cc) ---
(In reply to Luiz Von Dentz from comment #1)
> As far I know BE201 still use btusb, not really sure why a change to
> btintel_pcie would have any effect on a device over btusb but perhaps there
> is something changed that affect btintel which is shared between both.

I don't think that is correct. lspci describes the card as "00:14.7 Bluetooth
[0d11]: Intel Corporation Device [8086:a876] (rev 10)", Intel says it's an
CNVio card [1] (which I think is pcie only?), the problematic commit only
touches btintel_pcie.{c,h}, and btusb is not even loaded on my system (neither
before nor after the regression). 

[1]
https://www.intel.com/content/www/us/en/products/sku/230079/intel-wifi-7-be201/specifications.html

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

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