* network(networkd): treat bridge, bond, team and vlan links as wired
When the IP configuration lives on a bridge or link aggregate, the
enslaved member NIC reports opState=enslaved and the aggregate itself
was dropped by isWired() because only Type=ether was accepted, so the
widget showed Disconnected. Accept the same link types NetworkManager
already handles (#1581).
Also treat lxdbr*/incusbr* as virtual so container bridges, now of a
wired type, are not picked as the uplink.
Fixes#3463
* network(networkd): only treat a bridge as wired if it has a physical member
Name prefixes can't cover every container/VM bridge (lxcbr0, waydroid0,
...). Check the bridge's members in /sys/class/net/<br>/brif instead and
only accept it as the uplink when at least one member is not virtual
(veth*, tap*, ...). This replaces the lxdbr*/incusbr* prefixes.
(cherry picked from commit f886789835)
When iwd restarts it recreates wlan0 under a new D-Bus path. Only the
Station interface was tracked on InterfacesAdded, so devicePath kept
pointing at the dead object: the PropertiesChanged match for the new
device was never registered, updateState() read Powered from the old
path, and SetWiFiEnabled() failed with "No matching method found".
WiFiEnabled stayed false after handleStationRemoved() until DMS was
restarted.
Handle the Device interface on InterfacesAdded/InterfacesRemoved: move
devicePath and its signal match to the new object and take Name and
Powered from the signal. Process every interface in an
InterfacesRemoved signal instead of stopping at the first match.
Refs #3495
(cherry picked from commit ff50b492b1)
* feat: add cellular support
* fix(settings): resolve cellular tab index collision
* fix: address review feedback for cellular support
Bump the network API version, drop the deleted term freeze file,
regenerate the settings search index, and reuse existing translation
terms where they already cover the new cellular UI.
* fix: do not assign readonly userPreference when switching network type
Assigning SettingsData.networkPreference through the readonly alias
threw on click, so the Cellular segment in Control Center did nothing.
* fix: make Cellular selectable and the modem row clickable
The Control Center type switcher was bound only to preference, so
Cellular snapped back to Ethernet. The settings modem row highlighted
on hover but had no click handler. Show the modem as a connect target
when no APN profiles exist.
* fix: create a GSM profile when connecting cellular with none saved
Clicking Mobile broadband only activated an existing profile, so a
modem with no APN did nothing useful. Fall back to AddAndActivate
like ethernet connect.
* fix: address second round of cellular review feedback
- Guard cellularDevices map with devMutex via snapshot/by-iface/set/remove
helpers, mirroring the existing ethernet and wifi accessors
- Regenerate mocks with mockery v3 instead of hand-editing mock_Backend.go
- Reuse refreshCellularState() in SetCellularEnabled, DisconnectCellular,
DisconnectCellularDevice, ActivateCellularConnection and device removal
- Wrap retryErr rather than the first error when the auto-config retry fails
- Use Theme.surfaceTextSecondary/primaryHover/primaryHoverLight in the
cellular-off state instead of hand-rolled Qt.rgba values
- Drop the duplicate "No devices found" label so it can no longer render
twice with no modem and no saved config
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
* fix: address remaining cellular review comments
- Toast "Failed to toggle cellular" instead of always "enable"
- Remove unused enableCellularRadio() wrappers
- Store Control Center network type as a value, not a list index
* fix: pass nmcli connection selector as separate argument
`nmcli con mod uuid:<UUID>` is not valid syntax -- nmcli expects the
selector keyword as its own argument, so every priority update failed
with "no such connection profile". Because practically every NM profile
has a UUID, this silently disabled autoconnect-priority and route-metric
updates for ethernet and WiFi as well as cellular.
Also fix the Cellular settings status text, which said "Connecting..."
for a flag that only tracks radio enable/disable, and add the seven
cellular terms that were missing from the translation catalogs.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
* apply review comments
* fix lints
* update debug doc
* fix style inconsistencies
---------
Co-authored-by: Yechiel Worenklein <[email protected]>
Co-authored-by: Claude Opus 5 (1M context) <[email protected]>
Co-authored-by: bbedward <[email protected]>
* feat: add QR Generator modal with auto-generate on input debounce
- QRGeneratorModal with dual-buffer sequential fade animation
- Auto-generate QR code on typing stop (200ms debounce)
- Clear text button on input field
- Escape key closes modal
- Register modal in PopoutService and AppSearchService
- Go backend for text QR code generation
* fix: add QR Generator toggle to Launcher DMS settings
* qr: also add qrg built in launcher plugin
---------
Co-authored-by: bbedward <[email protected]>
* Add hotspot contract layer with capability-gated dispatch.
Establish the interface and manager plumbing for hotspot support so backends can opt in without expanding the core Backend contract.
Only backends that implement HotspotBackend get hotspot state propagated; others are forced to unsupported regardless of what they self-report.
* Add IPC handlers and API docs for hotspot actions.
Wire up configure/ start/ stop hotspot through the request router so the QML service layer can drive hotspot operations.
Bump API version to 27 and document the capability-gating contract clients should follow.
* Implement NetworkManager hotspot backend.
Implement HotspotBackend on NetworkManagerBackend with DMS-owned profile management, AP capability detection, and band validation.
Device resolution is deferred to StartHotspot when no device is specified, so profiles survive hardware changes.
* Isolate client Wi-Fi state from AP-mode connections.
Filter AP-mode profiles and access points out of all client Wi-Fi paths so the DMS hotspot (and user-created APs) never appear as saved networks, visible networks, or connected state.
This protects existing client behavior before hotspot controls are exposed in the UI.
* Prefer idle radios for automatic hotspot device selection.
* Add hotspot properties and methods to QML service layer.
Expose hotspot state and actions through DMSNetworkService, NetworkService, and LegacyNetworkService.
Capability is gated on API version and backend-reported support, not backend name checks, and stays stable when Wi-Fi radio is disabled.
* Add hotspot controls to Settings and Control Center.
Settings shows a hotspot setup card as a sibling in the Wi-Fi tab with SSID, password, device, band, save, and start/ stop controls.
Control Center shows a compact row that toggles a configured hotspot or routes to Settings for initial setup.
Both stay visible when Wi-Fi is disabled, explaining the requirement instead of hiding.
* Add translator context to hotspot strings.
* core: fix security and concurrency issues found in backend audit
Security:
- privesc: pipe the sudo password via stdin (sudo -S) instead of
embedding it in the command string, so it no longer appears in argv
(readable by any local user via /proc/<pid>/cmdline or ps)
- greeter: tokenize a session .desktop Exec= line into argv and execve
directly instead of running it through /bin/sh -c, closing a command-
injection path via user-writable ~/.local/share/wayland-sessions
- plugins: reject path-separator/.. in plugin id/name before joining
into a filesystem path, closing an arbitrary-directory-delete in the
uninstall/update fallback
- keybinds/hyprland: always quote unrecognized bind actions/keys when
writing generated Lua; only re-emit genuine round-tripped custom Lua
verbatim (tracked via an explicit flag), closing a Lua-injection path
- desktop/mimeapps: reject newline/bracket in mime/desktop-id fields so
they can't inject fake sections into the shared mimeapps.list
Robustness / concurrency:
- server: recover panics in the request-dispatch path so one bad
handler can't crash the daemon and drop every client
- go-wayland: recover panics in the shared dispatch choke point so a
malformed compositor event can't crash CLI tools / the daemon
- server: per-connection D-Bus client ID instead of a shared constant,
fixing cross-client signal delivery and subscription teardown
- network: guard the NetworkManager device maps with a mutex (a
concurrent map read/write here is an unrecoverable fatal error)
- cups: close the event channel on Stop() so Unsubscribe() of the last
subscriber no longer deadlocks; allocate the fresh channel in Start()
- freedesktop: reuse the shared session conn for the settings watcher
and tear it down in Close(), fixing a per-Manager conn+goroutine leak
- clipboard: mutex-guard lazy dbusConn creation
- geolocation: use WithMatchMember for the GeoClue2 LocationUpdated
signal (was WithMatchSender with an interface.member string, so the
match never fired and live location updates never arrived)
- screenshot: set failed=true on buffer/pool creation errors so the
dispatch loop doesn't wait forever for a ready/failed that never comes
* apply code review comments
---------
Co-authored-by: bbedward <[email protected]>
The networkd backend treated any link reporting Type=ether as a wired uplink.
Podman bridges and veth pairs report Type=ether, so they were classified as
ethernet: isWired() short-circuited on Type and never consulted looksVirtual(),
which also lacked a podman prefix.
The link map was also never pruned. Links discovered at enumeration or via
signals were kept forever, so torn-down container interfaces lingered as
routable and could win the wired-uplink slot over the real NIC -- leaving the
indicator showing WiFi while a wired connection was active and default-routed.
- isWired()/isWireless() exclude virtual interfaces before consulting Type, and
looksVirtual() now recognises podman.
- enumerateLinks() reconciles the cached map against ListLinks via syncLinks(),
pruning links that no longer appear so dead interfaces don't accumulate.
* networkd: classify links by Type instead of name prefix
The systemd-networkd backend decided wifi-vs-ethernet by checking
whether the interface name started with "wlan" or "wlp". Anything
else (that was not on a small virtual-prefix denylist) was treated
as wired ethernet. That misclassified two common cases:
* Nebula tunnels (kernel name like "nebula.homelab", Type=none,
Kind=tun) showed up as a wired ethernet device — DMS rendered an
"Ethernet connected" indicator whenever the overlay was up, even
with no physical NIC plugged in.
* Renamed wifi interfaces (e.g. systemd link files that rename
wlan0 to a friendlier name like "wifi") were also miscategorised
as ethernet, because they no longer matched wlan*/wlp*.
networkd already publishes the real link kind in the JSON returned
by the per-link Describe method ("ether", "wlan", "loopback",
"none"). Fetch it during enumerateLinks, cache it on linkInfo, and
classify against that. The old prefix logic is kept as a fallback
for the case where Describe ever fails to populate Type.
The package-level looksVirtual() helper replaces the unexported
isVirtualInterface method so the new classification helpers and
their tests can use it without needing a live backend.
Tests cover both the Type-based and fallback paths, including the
Nebula-shaped Type=none/tun case that motivated this change.
* networkd: cache linkType across signal ticks and unit-test Describe fallback
enumerateLinks runs on every PropertiesChanged signal under
/org/freedesktop/network1, which fires on carrier flap, DHCP renew, and
each address change. The previous version rebuilt every linkInfo from
scratch on each tick, including a synchronous Describe D-Bus round-trip
per link, despite the link Type being fixed at netlink creation. Preserve
existing entries when the D-Bus path matches, refreshing only ifindex,
and only call fetchLinkType on a genuinely new entry. A link torn down
and re-created at a different path still triggers a refetch.
Extract parseDescribeType from fetchLinkType so the JSON failure path —
malformed payload, missing Type field, wrong type for Type — can be
exercised without a live D-Bus connection. The classifier's fallback to
name-prefix heuristics already had coverage; this locks in that the seam
between Describe and the classifier surfaces an empty string on every
failure mode rather than misclassifying a link.
* fix: preserve pre-existing connection profiles on cancelled WiFi activation
When a pre-existing WiFi connection activation is cancelled, do not
remove the connection profile — only remove profiles for newly created
connections.
* feat: retrieve WiFi/VPN secrets from D-Bus secret service with unlock
Wireless, 802.1x, VPN, and WireGuard secrets are now looked up from
org.freedesktop.Secret before prompting the user. If the keyring is
locked, the unlock prompt is triggered via Prompt.Prompt() and the
lookup retried after the vault is unlocked.
- Added pre-run checks for greeter and setup commands to enforce policy restrictions
- Created cli-policy.default.json to define blocked commands and user messages for immutable environments.
* fix: add mockery v2 to nix flake pkgs
* feat: add requests for generating wifi qr codes as png files in /tmp, and to delete them later. only supports NetworkManager backend for now.
* feat: add modal for sharing wifi via qr code and saving the code as png file.
* fix: uncomment QR code file deletion
* network: light refactor and cleanup for QR code generation
---------
Co-authored-by: bbedward <[email protected]>
- we are dumb about importing by just trying to import everythting
- that caused errors to not be represented correctly
- just aggregate them all and present them in toast details
- Better would be to detect the type of file being imported, but this is
better than nothing
bbedward