This looks like a regression in kernel 5.18. The device is: Intel Corp. Wireless-AC 9260 Bluetooth Adapter After booting, my BT headphones connect to my host automatically. Once I suspend, this no longer happens, there's nothing in the system log: # bluetoothctl Agent registered AdvertisementMonitor path registered [ I'm switching on my BT headphones - no messages nothing, ok, switching them off ] [bluetooth]# power off Changing power off succeeded [CHG] Controller 48:89:E7:XX:YY:ZZ Powered: no [CHG] Controller 48:89:E7:XX:YY:ZZ Discovering: no [CHG] Controller 48:89:E7:XX:YY:ZZ Class: 0x00000000 [bluetooth]# power on [CHG] Controller 48:89:E7:XX:YY:ZZ Class: 0x007c0104 Changing power on succeeded [CHG] Controller 48:89:E7:XX:YY:ZZ Powered: yes [ I'm switching on my BT headphones again ] [CHG] Device 17:10:28:AA:BB:CC Connected: yes [NEW] Endpoint /org/bluez/hci0/dev_17_10_28_AA_BB_CC/sep1 [NEW] Endpoint /org/bluez/hci0/dev_17_10_28_AA_BB_CC/sep2 [NEW] Transport /org/bluez/hci0/dev_17_10_28_AA_BB_CC/fd1 [CHG] Transport /org/bluez/hci0/dev_17_10_28_AA_BB_CC/fd1 Volume: 0x000f (15) [DEL] Transport /org/bluez/hci0/dev_17_10_28_AA_BB_CC/fd1 [NEW] Transport /org/bluez/hci0/dev_17_10_28_AA_BB_CC/sep1/fd2 [CHG] Transport /org/bluez/hci0/dev_17_10_28_AA_BB_CC/sep1/fd2 Volume: 0x000f (15)
I'm looking at https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/log/drivers/bluetooth/btintel.c and there's nothing which could indicate a regression. Maybe it's an issue with bluez?
Also file here: https://github.com/bluez/bluez/issues/373
Downstream Fedora bug with some details and lots of people affected: https://bugzilla.redhat.com/show_bug.cgi?id=2100761
Must have been fixed by: https://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next.git/commit/?id=68253f3cd715e819bc4bff2b0e6b21234e259d56 Sadly, no one has pushed this very important fix to stable.