[mmaher88]

Summary

Logitech Options+ ships these product IDs for devices Logitune already supports, but our descriptors don't list them:

Descriptor	Added PID	Options+ entry
mx-master-3s	0xb043	model 2b043, a second "MX Master 3S" revision. Same depot, CIDs 0x52/0x53/0x56/0xC3/0xC4, DPI 200–8000 step 50
mx-master-3s-for-business	0xb044	model 2b044, same capabilities as b035
mx-master-2s	0x4069	Unifying WPID. Model 6b019 lists 046d_b019 and 046d_4069
mx-vertical	0xc08a	USB cable. Model eb020 lists 046d_407b, 046d_b020, 046d_c08a

Why it matters: DeviceManager::probeDevice only opens a session for a non-receiver hidraw node when findByPid() matches. A 3S reporting b043 over Bluetooth was never probed. Receiver-paired devices fall back to name matching, so they were less affected, but that fallback can't separate the 3S from the 3S For Business when the firmware name is ambiguous.

Wired MX Vertical (0xc08a): connection label is wrong

With this change, a cable-connected MX Vertical goes through probeDevice like this:

1. The HID++ report-ID check (0x85 0x11) passes on the HID++ interface.
2. 0xc08a is not a receiver PID, so it takes the direct branch, and findByPid(0xc08a) now matches.
3. deviceIndex = 0xFF (kDeviceIndexDirect). That's correct for a wired HID++ 2.0 device; Solaar uses 0xFF for both USB and Bluetooth direct connections.
4. connType is hard-coded to "Bluetooth" (DeviceManager.cpp, direct branch). This is wrong for USB. BatteryChip shows the Bluetooth icon and the Easy-Switch page says "Connected via Bluetooth" while the mouse is on its cable.

This PR doesn't fix the label. The fix needs the HID bus type. HIDIOCGRAWINFO already returns bustype (BUS_USB vs BUS_BLUETOOTH), but HidrawDevice::DeviceInfo drops it. It's a small follow-up: keep bustype and label the direct branch from it. Device identity isn't affected, because sessions merge by HID++ serial, not by PID or label.

0xc08a is only added to mx-vertical, which matches Options+ (it has no separate For Business model). A wired MX Vertical for Business would therefore resolve to the "MX Vertical" descriptor. The two descriptors differ only in hotspot positions.

The MX Vertical 0xFD button is intentionally left out. It needs hotspot coordinates.

Tests

• A DeviceSpec entry for each new PID (MX_Master_2S_4069, MX_Master_3S_b043, MX_Master_3S_For_Business_b044, MX_Vertical_c08a). Each one runs the full parameterized suite against its descriptor: controls, DPI, hotspots, gestures and features.
• The parameterized test names now end in the PID (…/MX_Master_3S_b034). Without that, two entries for one descriptor would produce duplicate gtest names.
• DeviceRegistry.FindByPidAndNameResolvesMxMaster3sRevisions: b034/b043 resolve to MX Master 3S and b035/b044 resolve to MX Master 3S For Business. That holds even when the reported name points at the sibling descriptor.
• I checked that the new cases fail when 0xb043 is removed from the descriptor (7 failures).

All of these pass locally on f6b1431: logitune-tests 836/836, logitune-tray-tests 9/9, logitune-qml-tests 72/72, generate-readme-devices.py --check (the table has no PID column, so README is unchanged), and generate_package_deps.py --check. Only descriptor JSON and tests change, so no hardware smoke test.

🤖 Generated with Claude Code