Description of problem: I have 2 laptops (identical F25 software) and the same mouse (Logitech MX Anywhere 2). The mouse works with the Broadcom controller (Apple laptop) and not the Qualcomm controller (Dell laptop). bluetoothctl does not see the mouse with Qualcomm although the hcidump shows identical packets with both. Version-Release number of selected component (if applicable): Fedora 25 bluez-libs-5.43-1.fc25.x86_64 bluez-5.43-1.fc25.x86_64 bluez-obexd-5.43-1.fc25.x86_64 bluez-cups-5.43-1.fc25.x86_64 How reproducible: 100% Steps to Reproduce: 1. Scan for new devices - mouse does not show up 2. 3. Actual results: Expected results: Additional info: Here is the info from the Apple computer (keep in mind Apple changes Vendor ID, but it is a Broadcom BCM20703A1 :: Apple Macbook Pro 11,5 :: - WORKS Bus 001 Device 002: ID 05ac:8290 Apple, Inc. Bluetooth Host Controller Device Descriptor: bLength 18 bDescriptorType 1 bcdUSB 2.01 bDeviceClass 239 Miscellaneous Device bDeviceSubClass 2 bDeviceProtocol 1 Interface Association bMaxPacketSize0 64 idVendor 0x05ac Apple, Inc. idProduct 0x8290 Bluetooth Host Controller bcdDevice 1.19 iManufacturer 1 Broadcom Corp. iProduct 2 Bluetooth USB Host Controller o hcidump > HCI Event: LE Meta Event (0x3e) plen 38 LE Advertising Report ADV_IND - Connectable undirected advertising (0) bdaddr CD:EE:9D:E9:B2:A2 (Random) Flags: 0x05 Unknown type 0x19 with 2 bytes data Complete service classes: 0x1812 Complete local name: 'MX Anywhere 2' RSSI: -58 > HCI Event: LE Meta Event (0x3e) plen 15 LE Advertising Report SCAN_RSP - Scan Response (4) bdaddr CD:EE:9D:E9:B2:A2 (Random) TX power level: 4 RSSI: -57 o bluetoothctl [bluetooth]# scan on Discovery started [CHG] Controller A0:99:9B:0B:CD:98 Discovering: yes [NEW] Device CD:EE:9D:E9:B2:A2 MX Anywhere 2 :: Dell XPS15 :: - NOT WORKING Bus 001 Device 003: ID 0cf3:e301 Atheros Communications, Inc. Device Descriptor: bLength 18 bDescriptorType 1 bcdUSB 2.01 bDeviceClass 224 Wireless bDeviceSubClass 1 Radio Frequency bDeviceProtocol 1 Bluetooth bMaxPacketSize0 64 idVendor 0x0cf3 Atheros Communications, Inc. idProduct 0xe301 o hcidump > HCI Event: LE Meta Event (0x3e) plen 38 LE Advertising Report ADV_IND - Connectable undirected advertising (0) bdaddr CD:EE:9D:E9:B2:A2 (Random) Flags: 0x05 Unknown type 0x19 with 2 bytes data Complete service classes: 0x1812 Complete local name: 'MX Anywhere 2' RSSI: -49 > HCI Event: LE Meta Event (0x3e) plen 15 LE Advertising Report SCAN_RSP - Scan Response (4) bdaddr CD:EE:9D:E9:B2:A2 (Random) TX power level: 4 RSSI: -50 o bluetoothctl [NEW] Controller 9C:B6:D0:D4:FE:74 localhost.localdomain [default] [NEW] Device 34:88:5D:3E:03:53 Logitech K810 [NEW] Device 04:88:E2:27:4A:F3 BeatsStudio Wireless [Logitech K810]# scan on Discovery started [CHG] Controller 9C:B6:D0:D4:FE:74 Discovering: yes [NEW] Device 00:04:4B:46:E8:02 SHIELD [NEW] Device C4:57:6E:36:D7:65 [TV]Samsung LED60 Mouse entry never makes it here. The only log in dmesg that I get when I start a scan on this device is: [Tue Jan 24 23:54:03 2017] hid-generic 0005:046D:B319.0008: unknown main item tag 0x0 Also note the way bluetoothctl changes the device to "[Logitech K810]# scan on" on the command line. The apple computer does not do this. The Logitech keyboard 810 you see is also a BTLE device, so I know the QCA driver can see those. nbot sure where to go from here. I have tried all Fedora kernels in F25, and like I said, both laptops have the exact software. Not sure if I need to post here, bluez, linux-wireless to get this resolved though. Let me know if I need to punt it elsewhere, or if more info is needed. Thanks, Greg Oliver
I got this mouse working using hcitool - I guess this Atheros adapter driver has some different default settings for BTLE devices than the Broadcom device which is weird since my Logitech K810 keyboard is also BTLE and it showed up with no changes on my Atheros controller. What worked to get it going was: hcitool lescan hcitool lewladd CD:EE:9D:E9:B2:A2 hcitool lerladd CD:EE:9D:E9:B2:A2 hcitool lecup (I do not know if the lewladd or lerladd would fix it independently - did not try). Then then the "scan on" by bluetoothctl immediately showed the device (as well as the gnome bluetooth app). I'm not sure where the issue should lie (bluez or atheros), but I am sure no one should need to do this to get a mouse working :) -Greg
I would like to note that the mouse takes a long time to re-acquire after the laptop sleeps goes idle for a while. Usually about 10 seconds. I also do not have that issue on the Broadcom controller.
Does this improve with blue 5.44? It had a number of BT-LE fixes
Hi Peter - Unfortunately not, but I am also not sure how I have 2 versions of the same apps / libs and not sure which it is using. I still had to run lerladd to get it going. bluez-5.42-2.fc25.x86_64 bluez-5.44-1.fc25.x86_64 bluez-cups-5.42-2.fc25.x86_64 bluez-cups-5.44-1.fc25.x86_64 bluez-libs-5.42-2.fc25.x86_64 bluez-libs-5.44-1.fc25.x86_64 bluez-obexd-5.42-2.fc25.x86_64 bluez-obexd-5.44-1.fc25.x86_64 gnome-bluetooth-3.20.0-2.fc25.x86_64 gnome-bluetooth-3.20.1-1.fc25.x86_64 gnome-bluetooth-libs-3.20.0-2.fc25.x86_64 gnome-bluetooth-libs-3.20.1-1.fc25.x86_64
One other thing to note - When I disable the wifi radio, it does not kill the bluetooth portion of the chip immediately, but if I let it sit idle or lock it - (after an undtermined amount of time), the bluetooth stops working completely until I re-enable the wifi.
Oh, BTW - just in case there are other cases - this laptop is a Dell XPS 1550 Product Name: XPS 15 9550 Product Name: 0N7TVV Adding hardware info as attachments..
Created attachment 1281304 [details] dmidecode
Created attachment 1281305 [details] lspci
Created attachment 1281306 [details] lsusb
I have this problem with a Dell XPS 13 using the atheros WiFi / bluetooth chip. I upgraded the firmware to get the wireless stable per this thread http://en.community.dell.com/techcenter/os-applications/f/4613/t/20002634 which ahead of the standard firmware. bluez just doesn't see the mouse on a scan but hcitool does.
This is working now in F26, fixed? Not sure when it started to work.
I suspect 5.46 probably did as there was a bunch of LE fixes in there