fix(client-linux,host): Deck video defaults to software decode + input-interception diagnostics

Video (Deck): the VAAPI zero-copy path renders corrupt/gray/washed-out on the
Deck — root-caused to Mesa >= 25.1 exporting radeonsi VCN decode surfaces TILED
(the Flatpak runtime's Mesa 26 drives both the decoder and GTK's GL, and GTK's
tiled-NV12 dmabuf import mishandles it; desktop Tier-1 validations ran distro
Mesa with linear export). `auto` now resolves to software on a Deck (clean,
correct-colour, easily handles 1280x800 HEVC); PUNKTFUNK_DECODER=vaapi still
forces the hw path, with the descriptor modifier dump + GSK_RENDERER as the
bisect levers. Also reserve extra_hw_frames=4 on the VAAPI decoder: the
presenter pins mapped surfaces past receive_frame, and the fixed pool recycling
a surface the renderer still samples is intermittent block corruption anywhere.

Input (Deck): with Steam Input ON for Punktfunk, SDL sees only Steam's virtual
X360 pad — the right trackpad arrives as a plain right stick and the left
trackpad/paddles/gyro not at all, silently. The client now checks once the
post-attach enumeration settles and raises a toast + warn naming the fix
(disable Steam Input for the shortcut). The host logs a one-shot warning when
InputPlumber is running (Bazzite default) since it can grab the virtual Deck
pad and re-emit it under a different identity.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

clients/linux/src/launch.rs

@@ -280,6 +280,35 @@ impl SessionUi {
         if self.app.fullscreen || self.app.settings.borrow().fullscreen_on_stream {
             self.app.window.fullscreen();
         }
+        // Steam Input owning the Deck's built-in controller is invisible degradation: SDL
+        // then sees only Steam's virtual X360 pad, so the right trackpad arrives at the
+        // host as a plain right stick (Steam's default template) and the left trackpad,
+        // paddles and gyro not at all. The real 28DE:1205 identity enumerates shortly
+        // after attach (the Valve HIDAPI drivers are session-scoped) — check once that
+        // settles and say so, instead of streaming silently degraded.
+        if crate::gamepad::is_steam_deck() {
+            let app = self.app.clone();
+            let stop = self.stop.clone();
+            glib::timeout_add_seconds_local_once(4, move || {
+                if stop.load(std::sync::atomic::Ordering::Relaxed) {
+                    return; // session already over
+                }
+                if app.gamepad.active().is_none_or(|pad| pad.steam_virtual) {
+                    tracing::warn!(
+                        "Steam Input is holding the built-in controller — trackpads, \
+                         paddles and gyro won't reach the host (right pad degrades to a \
+                         right stick). Disable Steam Input for Punktfunk to fix this."
+                    );
+                    let toast = adw::Toast::new(
+                        "Steam Input is holding the Deck's controls — trackpads, paddles \
+                         and gyro won't reach the game. Fix: Punktfunk → controller \
+                         settings → Disable Steam Input.",
+                    );
+                    toast.set_timeout(12);
+                    app.toasts.add_toast(toast);
+                }
+            });
+        }
         self.page = Some(p);
     }
 

clients/linux/src/video.rs

@@ -187,6 +187,25 @@ impl Decoder {
             .ok()
             .filter(|v| !v.is_empty())
             .unwrap_or_else(|| pref.to_string());
+        // The Steam Deck's VAAPI zero-copy path renders corrupt/gray/washed-out — validated live;
+        // software decode is clean, correct-colour, and the Deck's APU handles 1280×800 HEVC
+        // easily. Likely cause: since Mesa 25.1 radeonsi exports VCN decode surfaces TILED (with
+        // AMD modifiers) instead of linear, and inside the Flatpak both the VAAPI driver and GTK's
+        // GL come from the runtime's Mesa 26.x — GTK's tiled-NV12 dmabuf import mishandles the new
+        // layout (desktop AMD/Intel boxes validated Tier-1 ran distro Mesa with linear export).
+        // So `auto` resolves to software on a Deck; an explicit `vaapi` (Settings or
+        // PUNKTFUNK_DECODER=vaapi) still forces the hw path for testing — the first-frame
+        // descriptor dump logs the modifier (LINEAR = 0x0), and GSK_RENDERER=ngl|vulkan bisects
+        // the import side.
+        let choice = if (choice == "auto" || choice.is_empty()) && crate::gamepad::is_steam_deck() {
+            tracing::info!(
+                "Steam Deck — defaulting to software decode (AMD VAAPI dmabuf is broken on this \
+                 SteamOS+Mesa combo); set the decoder to `vaapi` to override"
+            );
+            "software".to_string()
+        } else {
+            choice
+        };
         if choice != "software" {
             match VaapiDecoder::new(codec_id) {
                 Ok(v) => {
@@ -456,6 +475,14 @@ impl VaapiDecoder {
             (*ctx).get_format = Some(pick_vaapi);
             (*ctx).flags |= ffi::AV_CODEC_FLAG_LOW_DELAY as i32;
             (*ctx).thread_count = 1; // hwaccel: threads only add latency
+
+            // The presenter holds mapped surfaces PAST receive_frame (the paintable's
+            // current texture + the newest frame in flight each pin one until GDK's
+            // release func) — surfaces libavcodec doesn't know are missing from its
+            // fixed-size VAAPI pool. Without headroom the decoder can recycle a surface
+            // the renderer is still sampling (intermittent block corruption) or fail
+            // allocation under scheduling jitter.
+            (*ctx).extra_hw_frames = 4;
             let r = ffi::avcodec_open2(ctx, codec, ptr::null_mut());
             if r < 0 {
                 let mut ctx = ctx;