Logi POP Mouse (BLE-only) and K380 keyboard: consistent "Device timed out" on device add

[daramosn]

Description
Both my Logi POP Mouse and Bluetooth Multi-Device Keyboard K380 are detected
by logid (button/gesture capabilities are correctly enumerated), but device
initialization always fails with "Device timed out". This happens 100% of
the time, on every launch/restart, not intermittently.

Both devices are connected via plain Bluetooth (BLE), with no USB
Unifying/Bolt receiver involved.

Environment

• OS: Pop!_OS 24.04
• Kernel: 7.0.11-76070011-generic
• logid version: v0.3.3 (both the Ubuntu-packaged build and a fresh build
  from the current PixlOne/logiops master branch show identical behavior)
• BlueZ: 5.72
• Devices:
  • Logi POP Mouse (HID 046D:B030, BLE)
  • Keyboard K380 (HID 046D:B342, BLE)

Steps to reproduce

1. Pair/connect Logi POP Mouse and K380 over Bluetooth (no receiver)
2. sudo systemctl restart logid
3. journalctl -u logid -f

Output
[INFO] Device found: Logi POP Mouse on /dev/hidraw0:255
[DEBUG] /dev/hidraw0:255 remappable buttons:
[DEBUG] CID | reprog? | fn key? | mouse key? | gesture support?
[DEBUG] 0x50 | | | YES |
[DEBUG] 0x51 | | | YES |
[DEBUG] 0x52 | YES | | YES | YES
[DEBUG] 0xd7 | YES | | | YES
[DEBUG] 0x108 | YES | | YES | YES
[WARN] Error adding device /dev/hidraw0: Device timed out

[INFO] Device found: Bluetooth Multi-Device Keyboard K380 on /dev/hidraw1:255
[DEBUG] /dev/hidraw1:255 remappable buttons:
[DEBUG] CID | reprog? | fn key? | mouse key? | gesture support?
... (all reprog buttons enumerated normally) ...
[WARN] Error adding device /dev/hidraw1: Device timed out

What I've already ruled out

• Solaar was running simultaneously -> stopped and uninstalled it, no change
• Restarted bluetooth.service before logid -> no change
• Restarted logid multiple times, with devices freshly woken/moved -> no change
• Confirmed hid_logitech_hidpp kernel module is loaded, but neither device's
  PID (046D:B030 / 046D:B342) is in its supported alias table
  (modinfo hid_logitech_hidpp | grep alias), so both fall back to
  hid-generic in the kernel, which logs:
  "Event data for report 3 was too short (7 vs 1)"
  repeatedly for both devices.
• Rebuilt logid from the current master branch (still reports as v0.3.3) ->
  identical timeout behavior
• Confirmed via Solaar that the mouse does respond correctly to HID++ over
  this same Bluetooth connection (full feature table read successfully,
  including REPROG CONTROLS V4 with the same CIDs logid reports:
  82(0x52)/264(0x108))

Given that Solaar can complete a full HID++ handshake with the same device
over the same Bluetooth connection, this looks like it might be related to
BLE-only Logitech devices requiring "long report only" HID++ framing
(no short reports supported) — similar to what upstream kernel commits like
"HID: logitech: Support HID++ devices without short reports" had to
specifically account for in hid-logitech-hidpp. Since logid talks to
/dev/hidraw directly, it may not be applying that fallback itself. I have
not looked at the logid source in enough depth to confirm this, but wanted
to share the diagnostic trail in case it's useful.

Happy to test patches or provide more verbose logs if helpful.