On Fri, Sep 08, 2023 at 05:02:54PM +0200, Antoine Nauzet wrote: > Le ven. 8 sept. 2023 à 07:55, Sean Greenslade <sean@xxxxxxxxxxxxxxxxxx> a > > I would start with using your distro's pulse package and checking both > > the system logs and the pulseaudio logs. I would expect there to be some > > chatter when a bluetooth device event like that happens. > > > > --Sean > > > > > I'm back to distro's pulse package. When the headset_audio_gateway > capability is switch to not available, the pulse audio logs this: > > ( 543.481| 8.707) D: [pulseaudio] bluez5-util.c: Device > /org/bluez/hci0/dev_78_29_C9_4F_C2_4E removed This line only happens when pulse receives an "InterfacesRemoved" message from Bluez. Whatever is going wrong in your system, it's happening in the bluetooth stack itself. Pulse has no control over it. I think you'll need to ask for help from the Bluez project. > ( 543.482| 0.000) D: [pulseaudio] backend-ofono.c: HF card removed: > /card_2 > ( 543.482| 0.000) D: [pulseaudio] bluez5-util.c: Transport /card_2 state: > idle -> disconnected > ( 543.482| 0.000) D: [pulseaudio] card.c: Setting card > bluez_card.4C_DD_31_43_8E_FF profile headset_audio_gateway to availability > status no > ( 543.482| 0.000) D: [pulseaudio] device-port.c: Setting port > phone-output to status no > ( 543.482| 0.000) D: [pulseaudio] core-subscribe.c: Dropped redundant > event due to change event. > > Is "card_2" the card "2" display when typing pactl list cards? No, /card_2 is a reference to the transport path from the bluetooth stack. Nothing to do with the card numberings in pulse. --Sean