[Bug 287813] USB headset not properly initialized

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=287813

--- Comment #6 from J.R. Oldroyd <fbsd@opal.com> ---
Let's not forget that the "pactl exit" hack makes the device work perfectly. 
After that, I can play multiple audios using mplayer and I can have hours-long
Meet/Zoom calls in firefox as well as watching YouTube vids (albeit after each
firefox close, I need to rerun the "pactl exit" hack).

The fact that it works seems to suggest that we do not have a device problem or
a firmware problem.  This points most definitely to a driver/device
initialization problem.

To answer Tomasz's questions:

The USB-Bluetooth dongle identifies in Linux's lsusb (see [1]) as a GN Netcom
model BTV5.2.  FreeBSD's usbconfig shows:
        ugen0.3: <vendor 0x0b0e BTV5.2> at usbus0
The Jabra Link 380a MS appears to have the same USB vendor and model IDs.

For completeness, the Bluetooth headset it came with is a Mairdi M200V.  Xiamen
Mairdi Electronic Technology Co Ltd.  The headset works perfectly when paired
to a phone's Bluetooth.

Connecting the dongle directly to USB port, no cable.  Laptop use.  Connector
is a USB-A tried in both USB 2.0 and 3.0 ports.  Same results for both.

Tried in Linux (same problem) but not in MacOS or Windows (don't currently have
those available).  But I note again that the device DOES work on FreeBSD after
running the "pactl exit" hack as described in the OP.  A special driver or
application is clearly not needed.

Not tried upgrading firmware of the device.  Would not appear to be necessary
as the device can be made to work with the hack.  There is firmware on the
Jabra support page - latest version is from 2022/10/25, three years ago.  MacOS
or Windows are required to check and update firmware - neither currently
available but I will check later.  The device was new a few months ago, so
likely it does have firmware from 2022.

As I said, I am suspecting that this may be a case of the driver not properly
initializing its own parameters or those of the device.  With the D50488 patch
reverted - i.e., the driver close routine is broken, pulseaudio is able to open
the driver, set driver parameters, perhaps talk to the device and set its
parameters, and then go away but, because the close code is broken, those
settings remain and the device then works perfectly.  With the D50488 patch in
place, the close routine has been fixed, so when pulseaudio goes away, whatever
settings were done are now lost and it no longer works.

[1] Info from the Linux command "lsusb -v" for this device:
    https://opal.com/~jr/btv5.2/linux_lsusb.txt

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