Re: Unable to get Bluetooth to work on main with CSR5810

On Wed, 4 Mar 2026, Kevin Oberman wrote:

Hi Kevin,
Cc: bluetooth mailing list as well


my 2 uneducated cts from the little BT I have learnt in the last years.


> Here is the complete USB content from dmesg -a:
> ugen1.4: <vendor 0x0a12 CSR8510 A10> at usbus1
> ubt1 on uhub0
> ubt1: <vendor 0x0a12 CSR8510 A10, class 224/1, rev 2.00/88.91, addr 3> on
> usbus1
> ng_hci_process_command_timeout: ubt1hci - unable to complete HCI command
> OGF=0x3, OCF=0x3. Timeout
> ng_hci_process_command_timeout: ubt0hci - unable to complete HCI command
> OGF=0x3, OCF=0x3. Timeout

Hmm, that's a baseband reset if I remember this right.

> ubt1: ubt_bulk_read_callback:1165: bulk-in transfer failed: USB_ERR_IOERROR
> ubt1: ubt_intr_read_callback:1066: interrupt transfer failed:
> USB_ERR_IOERROR

I assume net.bluetooth.usb_isoc_enable="0" will not help as this is too early.
[ See https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=274707 ]

> ugen1.4: <vendor 0x0a12 CSR8510 A10> at usbus1 (disconnected)
> ubt1: at uhub0, port 1, addr 3 (disconnected)
> ubt1: detached
>
> The last 3 lines are from unplugging the fob.

I really just wonder if you got a cheap clone;  there's been issues
with that for CSR devices and they would need detection/quirks for setup.

We need a usbconfig -d 1.4 dump_all_desc for the device (1.4 may change for the ugen)
and a dump of NG_HCI_OCF_READ_LOCAL_VER to validate that but I dont think a
hccontrol read_local_version_information would work at this point
so we could patch it for that.

If someone will do that it would be great but for me it would be
easier to fix the CNVIo issues on Intel I think.

/bz

-- 
Bjoern A. Zeeb                                                     r15:7