Hi,
Not sure if this is the right place, but I've noticed excessive logging
coming from the bluetooth stack while using the QCA WCN6855 combined
bluetooth/wifi card. Using a bluetooth mouse results in a few messages
per minute. Using wireless headphones results in upwards of 3 messages
per second.
The computer is a Lenovo T14 Amd Gen 2 (sku 20XK006BUS) Direct link
here: https://tinyurl.com/5xrad7w5
The software is Debian Stable, with kernel versions of 5.18, 5.19, and 6.0.
The excessive logging is in the the form of:
[3400058.603040] Bluetooth: hci0: ACL packet for unknown connection
handle 3804
[3400058.603396] Bluetooth: hci0: ACL packet for unknown connection
handle 3804
[3400058.603916] Bluetooth: hci0: ACL packet for unknown connection
handle 3804
[3400059.000540] Bluetooth: hci0: ACL packet for unknown connection
handle 3804
[3400061.130513] Bluetooth: hci0: ACL packet for unknown connection
handle 3804
[3400083.158218] Bluetooth: hci0: ACL packet for unknown connection
handle 3804
[3400083.158622] Bluetooth: hci0: ACL packet for unknown connection
handle 3804
[3400083.159139] Bluetooth: hci0: ACL packet for unknown connection
handle 3804
[3400083.623400] Bluetooth: hci0: ACL packet for unknown connection
handle 3804
"Connection handle 3804" coincides with this commit
(https://github.com/torvalds/linux/commit/c614ca3f7476934de54dd731e09d094ad822696c)
from Oct 2018 where extra logic was added to hci_qca.c to handle the
fact that QCA devices use ACL packets to transmit extra controller
debugging info. The WCN6855 chipset does not use the UART interface but
uses btusb.c. The logic from hci_qca.c was never backported for the USB
interface though. I believe that is the problem. There are other
manufactures doing similar within btusb.c, see btusb_recv_acl_mtk().
It also seems at one time QCA was disabling the extra logging via some
extra commands, at least for the devices that use btqca.c:
https://github.com/torvalds/linux/commit/83e81961ff7ef75f97756f316caea5aa6bcc19cc
So because this is all QCA code, I'm not sure if I'm at liberty to
provide a patch. For now, I've simply built the bluetooth stack with
logging for "unknown connection handle" disabled (within
hci_recv_frame() ). But could someone in the know perhaps forward this
to any applicable maintainers?
Feel free to ask me for any extra info if needed.
Kind regards,
Samuel Smith
