Bug 2375838 - Logitech MX Master 3 won't reconnect on reboot
Summary: Logitech MX Master 3 won't reconnect on reboot
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Fedora
Classification: Fedora
Component: bluez
Version: 42
Hardware: x86_64
OS: Linux
unspecified
medium
Target Milestone: ---
Assignee: Peter Robinson
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2025-07-02 03:42 UTC by Berend De Schouwer
Modified: 2025-07-14 15:36 UTC (History)
5 users (show)

Fixed In Version: bluez-5.83-2.fc42 bluez-5.83-2.fc41
Clone Of:
Environment:
Last Closed: 2025-07-12 01:45:56 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
lsusb -v of the USB controller (12.92 KB, text/plain)
2025-07-02 05:31 UTC, Berend De Schouwer
no flags Details
bluetoothctl show controller (2.21 KB, text/plain)
2025-07-02 13:57 UTC, Berend De Schouwer
no flags Details
hcidump output (58.17 KB, text/plain)
2025-07-03 07:26 UTC, Berend De Schouwer
no flags Details

Description Berend De Schouwer 2025-07-02 03:42:56 UTC
When booting or rebooting, a bluetooth Logitech MX Master 3 will no longer (re-)connect to the laptop.  Other (non-logitech) bluetooth devices connect fine, and the Logitech mouse will connect to other laptops running older Linux.

I have tried powering the mouse on and off, using the connect button on the mouse, and connecting using gnome-settings.  I have also tried restarting the bluetooth service.

Either of the following allows reconnecting:
* forgetting the bluetooth device
* rmmod hid_logitech_hidpp (it reloads automatically)

In the past, it was sometimes slow to connect, see bug #2323758, but as of recently it just won't reconnect at all.

Messages from bluetooth.service and the kernel attached.  I'm going to assume it could be either, since both were updated in the last month.

It *seems* to be much more prominent when the laptop reconnects quickly, eg. when doing software updates, and dracut includes bluetooth to unlock luks.

I can try, but would appreciate guidance:
* add -d to bluetoothd
* add a parameter to the bluetooth module (which one?)
* edit /etc/bluetooth/main.conf, candidates
  * FastConnectable
  * ResumeDelay
  * Others?
* run hcidump automatically on reboot, not sure what I'm looking for.
* rmmod hid_logitech automatically
* restart bluetooth on suspend/wakeup (this crashes wireplumber)

Reproducible: Sometimes

Steps to Reproduce:
1. Have logitech mx master 3 paired
2. suspend/wakeup or reboot
3. move mouse
Actual Results:
mouse does not connect to bluetooth

Expected Results:
mouse reconnects to bluetooth

Additional Information:
Kernel:

[78413.209833] Bluetooth: hci0: HW/SW Version: 0x008a008a, Build Time: 20250523111333
[78413.351318] Bluetooth: hci0: Device setup in 162865 usecs
[78413.351324] Bluetooth: hci0: HCI Enhanced Setup Synchronous Connection command is advertised, but not supported.
[78413.384780] Bluetooth: hci0: AOSP extensions version v1.00
[78413.384785] Bluetooth: hci0: AOSP quality report is supported
[78417.405858] Bluetooth: hci0: ACL packet for unknown connection handle 3837

bluetooth service:

Jul 02 04:49:59 cijfer.deschouwer.co.za bluetoothd[35650]: profiles/deviceinfo/deviceinfo.c:read_pnpid_cb() Error reading PNP_ID value: Request attribute has encountered an unlikely error
Jul 02 04:49:59 cijfer.deschouwer.co.za bluetoothd[35650]: profiles/input/hog-lib.c:proto_mode_read_cb() Protocol Mode characteristic read failed: Request attribute has encountered an unlikely error
Jul 02 04:49:59 cijfer.deschouwer.co.za bluetoothd[35650]: profiles/input/hog-lib.c:report_read_cb() Error reading Report value: Request attribute has encountered an unlikely error
Jul 02 04:49:59 cijfer.deschouwer.co.za bluetoothd[35650]: profiles/input/hog-lib.c:report_reference_cb() Read Report Reference descriptor failed: Request attribute has encountered an unlikely error
Jul 02 04:49:59 cijfer.deschouwer.co.za bluetoothd[35650]: profiles/input/hog-lib.c:report_read_cb() Error reading Report value: Request attribute has encountered an unlikely error
... loops ...

Comment 1 Berend De Schouwer 2025-07-02 05:31:17 UTC
Created attachment 2095867 [details]
lsusb -v of the USB controller

Comment 2 Berend De Schouwer 2025-07-02 05:32:02 UTC
More info:

Bluetooth service logs from a good connection:

Jul 02 07:04:10 cijfer.deschouwer.co.za bluetoothd[46093]: src/gatt-database.c:connect_cb() New incoming LE ATT connection
Jul 02 07:04:10 cijfer.deschouwer.co.za bluetoothd[46093]: src/device.c:device_attach_att() Elevating security level since LTK is available
Jul 02 07:04:10 cijfer.deschouwer.co.za bluetoothd[46093]: attrib/gattrib.c:g_attrib_ref() 0x5646ea876a80: g_attrib_ref=1
Jul 02 07:04:10 cijfer.deschouwer.co.za bluetoothd[46093]: src/device.c:load_gatt_db() Restoring F3:BF:2E:15:16:07 gatt database from file


Logitech MX Master 3 is an LE device at randomised F3:BF:2E:15:16:07, from bluetoothctl info ...

Device F3:BF:2E:15:16:07 (random)
	Name: MX Master 3
	Alias: MX Master 3
	Appearance: 0x03c2 (962)
	Icon: input-mouse
	Paired: yes
	Bonded: yes
	Trusted: yes
	Blocked: no
	Connected: yes
	WakeAllowed: yes
	LegacyPairing: no
	CablePairing: no
	UUID: Generic Access Profile    (00001800-0000-1000-8000-00805f9b34fb)
	UUID: Generic Attribute Profile (00001801-0000-1000-8000-00805f9b34fb)
	UUID: Device Information        (0000180a-0000-1000-8000-00805f9b34fb)
	UUID: Battery Service           (0000180f-0000-1000-8000-00805f9b34fb)
	UUID: Human Interface Device    (00001812-0000-1000-8000-00805f9b34fb)
	UUID: Vendor specific           (00010000-0000-1000-8000-011f2000046d)
	Modalias: usb:v046DpB023d0015
	Battery Percentage: 0x32 (50)


Controller is 14:13:33:55:5E:F2, a USB (but inside laptop) Mediatek, from hciconfig -a

hci0:	Type: Primary  Bus: USB
	BD Address: 14:13:33:55:5E:F2  ACL MTU: 1021:6  SCO MTU: 240:8
	UP RUNNING PSCAN ISCAN INQUIRY 
	RX bytes:3483120 acl:107703 sco:0 events:123998 errors:0
	TX bytes:81567809 acl:96238 sco:0 commands:12351 errors:0
	Features: 0xbf 0x3e 0x8d 0xfe 0xdb 0xff 0x7b 0x87
	Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3 
	Link policy: RSWITCH SNIFF 
	Link mode: PERIPHERAL ACCEPT 
	Name: 'cijfer.deschouwer.co.za'
	Class: 0x7c010c
	Service Classes: Rendering, Capturing, Object Transfer, Audio, Telephony
	Device Class: Computer, Laptop
	HCI Version: 5.2 (0xb)  Revision: 0x615
	LMP Version: 5.2 (0xb)  Subversion: 0x2411
	Manufacturer: MediaTek, Inc. (70)


Wifi is on PCI.

Mouse shows up in HID as:

Path         : /dev/hidraw4
Index        : 255
Product ID   : 046d:B023
Protocol     : HID++ 4.5
Unit ID      : B246ED89
  1          : BOT 95.01.B0015
  0          : MPM 19.01.B0015
Notifications: (none)


Bluez version: bluez-5.83-1.fc42.x86_64
Kernel version: kernel-6.15.4-200.fc42.x86_64

Comment 3 Peter Robinson 2025-07-02 08:29:09 UTC
> Bluez version: bluez-5.83-1.fc42.x86_64
> Kernel version: kernel-6.15.4-200.fc42.x86_64

What version of linux-firmware?

I think it's a mediatek controller, but I couldn't see what model, can you provide those details.

When did the problem start? If you downgrade bluez does it start working again? 'dnf downgrade bluez' will likely give you 5.80, it would be useful to test bluez-5.81-2.fc42 but koji is down ATM so might not be easy to obtain.

Comment 4 Berend De Schouwer 2025-07-02 13:57:08 UTC
(In reply to Peter Robinson from comment #3)
> > Bluez version: bluez-5.83-1.fc42.x86_64
> > Kernel version: kernel-6.15.4-200.fc42.x86_64
> 
> What version of linux-firmware?

linux-firmware-20250627-1.fc42.noarch

> I think it's a mediatek controller, but I couldn't see what model, can you
> provide those details.

It is a Mediatek controller.  I'm sorry I didn't include that in the original
report.  It's USB, included in the laptop.

So vendor:product is 13d3:3563.  I haven't found much of "product".

ID 13d3:3563 IMC Networks Wireless_Device
  iManufacturer           5 MediaTek Inc.
  iProduct                6 Wireless_Device
  iSerial                 7 000000000
Bluetooth: hci0: HW/SW Version: 0x008a008a, Build Time: 20250523111333

/sys/bus/usb/... gives version 2.10.

lshw gives:
                 *-usb
                      description: Bluetooth wireless interface
                      product: Wireless_Device
                      vendor: MediaTek Inc.
                      physical id: 3
                      bus info: usb@3:3
                      version: 1.00
                      serial: 000000000
                      capabilities: usb-2.10 bluetooth
                      configuration: driver=btusb maxpower=100mA speed=480Mbit/s


Controller info from bluetoothctl attached.

> When did the problem start?

About a week or two ago; which means updates from up to a month ago.

> If you downgrade bluez does it start working
> again? 'dnf downgrade bluez' will likely give you 5.80, it would be useful
> to test bluez-5.81-2.fc42 but koji is down ATM so might not be easy to
> obtain.

Will bisect.

For now, I've downgraded to 5.80-1.fc42; and I'll test.

Comment 5 Berend De Schouwer 2025-07-02 13:57:56 UTC
Created attachment 2095965 [details]
bluetoothctl show controller

Comment 6 Berend De Schouwer 2025-07-03 07:26:08 UTC
Created attachment 2096004 [details]
hcidump output

hcidump output.  The mouse connects more-or-less where you first find "handle 512"

Comment 7 Berend De Schouwer 2025-07-03 07:31:12 UTC
With bluez-5.80-1.fc42.x86_64 the mouse took ± 10 minutes to connect today, without the laptop suspend/reboot-ing.  I was in a video meeting, and didn't use the mouse for an hour.

The bluetooth keyboard (Dell) connected immediately, the mouse took 10 actual minutes.  I've attached hcidump output.  The mouse connected more-or-less where you see "handle 512" for the first time.  I have no idea if this helps, or if other bluetoothctl or hcidump options would help.

I spent the first few minutes switching off other bluetooth devices, only to realise I can see the neighbours TV anyway.  I'm planning to setup rotating files for raw dumps -- again, don't know if this helps anyone.

Comment 8 Peter Robinson 2025-07-03 07:35:54 UTC
(In reply to Berend De Schouwer from comment #7)
> With bluez-5.80-1.fc42.x86_64 the mouse took ± 10 minutes to connect today,
> without the laptop suspend/reboot-ing.  I was in a video meeting, and didn't
> use the mouse for an hour.

So it's not bluez that's the regression here then, can you try and work out whether it was firmware or kernel that regressed it?

This damn mouse has been problematic on and off for years :-/

Comment 9 Berend De Schouwer 2025-07-03 08:04:29 UTC
Downgraded to linux-firmware-0:20250311-1.fc42.noarch.  I'm not sure about the mouse firmware.  There doesn't appear to be an update for this mouse.

The oldest kernel I still have installed is kernel-6.14.11-300.fc42.x86_64, which the changelog says comes from 10 June.  After that it's Koji.

I've got some more hcidumps.  Still looking for raw bluetooth dumps.


My guess, after the video meeting incident, is it's power saving or a bad re-randomization of the MAC, which is now F3:BF:2E:15:16:08 -- it used to be F3:BF:2E:15:16:07 before today.

Comment 10 Peter Robinson 2025-07-09 09:15:08 UTC
So the linux-firmware-20250708-1.fc42 has a "mediatek MT7921: update bluetooth firmware to 20250625154126" update and I'm pushing a new bluez build bluez-5.83-2.fc42 build which fixes a bluez issue with cmdline (I'm not sure this will help this). So let's clutch at some straws and see if either of those help!

Comment 11 Fedora Update System 2025-07-09 09:43:23 UTC
FEDORA-2025-c10d623d39 (bluez-5.83-2.fc42) has been submitted as an update to Fedora 42.
https://bodhi.fedoraproject.org/updates/FEDORA-2025-c10d623d39

Comment 12 Fedora Update System 2025-07-09 09:43:29 UTC
FEDORA-2025-e53a11a6d0 (bluez-5.83-2.fc41) has been submitted as an update to Fedora 41.
https://bodhi.fedoraproject.org/updates/FEDORA-2025-e53a11a6d0

Comment 13 Berend De Schouwer 2025-07-10 04:04:03 UTC
(In reply to Peter Robinson from comment #10)
> So the linux-firmware-20250708-1.fc42 has a "mediatek MT7921: update
> bluetooth firmware to 20250625154126" update and I'm pushing a new bluez
> build bluez-5.83-2.fc42 build which fixes a bluez issue with cmdline (I'm
> not sure this will help this). So let's clutch at some straws and see if
> either of those help!

Testing

Comment 14 Berend De Schouwer 2025-07-10 04:05:17 UTC
(In reply to Berend De Schouwer from comment #9)
> Downgraded to linux-firmware-0:20250311-1.fc42.noarch.

It turns out downgrading linux-firmware did not downgrade mt7xxx-firmware, so the previous testing may be invalid.

Comment 15 Fedora Update System 2025-07-10 16:01:31 UTC
FEDORA-2025-c10d623d39 has been pushed to the Fedora 42 testing repository.
Soon you'll be able to install the update with the following command:
`sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2025-c10d623d39`
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2025-c10d623d39

See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.

Comment 16 Fedora Update System 2025-07-10 18:45:57 UTC
FEDORA-2025-e53a11a6d0 has been pushed to the Fedora 41 testing repository.
Soon you'll be able to install the update with the following command:
`sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2025-e53a11a6d0`
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2025-e53a11a6d0

See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.

Comment 17 Fedora Update System 2025-07-12 01:45:56 UTC
FEDORA-2025-c10d623d39 (bluez-5.83-2.fc42) has been pushed to the Fedora 42 stable repository.
If problem still persists, please make note of it in this bug report.

Comment 18 Fedora Update System 2025-07-13 02:58:18 UTC
FEDORA-2025-e53a11a6d0 (bluez-5.83-2.fc41) has been pushed to the Fedora 41 stable repository.
If problem still persists, please make note of it in this bug report.

Comment 19 Berend De Schouwer 2025-07-14 15:36:49 UTC
It happened again this morning.

The good news is I have both tcpdump and debug logs of bluetooth. (can append)

The bad news is it looks like the controller filters out the mouse.

Maybe if I can disable own address randomization, or change Scan Filter Policy: Accept all advertisements, except directed advertisements not addressed to this device (0x00)


Note You need to log in before you can comment on or make changes to this bug.