Networking-audit deferred plan §2. The host↔client offset was measured once
at connect; an NTP step or slow drift silently corrupted the clock-based
jump-to-live signal, the ABR one-way-delay signal, and every latency stat —
4a3b1ae2's disarm backstop stopped the IDR storm but lost the detector for
the session. Now the client re-estimates mid-stream and recovers it.
- quic: ClockResync — the connect-time 8-round probe/echo estimate as a
select!-driven state machine (rounds matched by echoed t1, stale batches
ignored), plus accept_resync (batch min-RTT ≤ max(2 ms, 1.5× connect RTT)
so a congested window can never bias the offset). No wire change: the
host has always answered ClockProbe at any time on the control stream.
- client: the offset lives in an Arc<AtomicI64> seeded at connect; the
control task re-probes every 60 s and immediately after the pump's FIRST
no-op clock flush (the "clock stepped under me" signal, sent on the next
report tick). On apply: store, reset stale_frames/noop_clock_flushes,
re-arm the clock detector if a step had disarmed it. The disarm heuristic
stays as the final backstop. Public NativeClient::clock_offset_ns keeps
the connect-time value (ABI untouched); new clock_offset_now_ns() /
clock_offset_shared() expose the live value.
- consumers migrated to the live offset: pf-client-core session stats, the
pf-presenter e2e stamp, Windows session/render, Android feeder/drain/
DisplayTracker (the tracker holds the shared handle, not the client, so
the leaked render-callback refcount can't pin the session).
- probe: --clock-resync runs a second full handshake mid-connection and
asserts a sane, consistent estimate. Live against the local canary host:
offsets 8646/2139 ns, disagreement 6 µs, 8/8 rounds — OK.
Unit tests cover the round collection, stale-echo rejection, batch restart,
min-RTT selection, and the acceptance guard. cargo ndk check green.
Remaining manual validation: `sudo date -s "+2 sec"` on a live streaming
client → expect one no-op flush, a re-sync, re-armed detector, no IDR pulse.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
One stat model everywhere (design/stats-unification.md): four measurement
points (capture/received/decoded/displayed), three stages that tile the
interval exactly, and a HUD that shows the addition explicitly —
end-to-end 14.2 ms p50 · 19.8 p95 · capture→on-glass
= host+network 9.8 + decode 2.1 + display 2.3
replacing each client's ad-hoc mix of overlapping absolutes (the Apple HUD's
three arrow lines that looked sequential but weren't), mean-vs-median decode
times (Windows/Linux), missing same-host-clock flags (Windows/Linux), and
three different names for the same capture→received measurement (probe's
"reassembled", Apple/Android's "client", Windows/Linux's post-decode "lat").
Per client: Apple threads receivedNs through the VT decode via the frame
refcon bit pattern so the decode stage exists at all (stage-1 fallback
honestly degrades to a capture→received headline); Windows carries
FrameTimes through the existing frame channel to the render thread and adds
e2e p50/p95 post-Present; Linux stamps received at AU pop and rides
decoded_ns on DecodedFrame to the paintable-set site; Android pairs receipt
stamps with MediaCodec output buffers via the codec's pts round-trip (JNI
stats array 14→16 doubles, indexes 0-13 unchanged). fps now uniformly counts
received AUs; lost/(received+lost) per window, hidden at zero.
docs-site gains "Understanding the Stats Overlay": what each line means, why
the equation only approximately sums (percentiles), and a line-by-line
Moonlight/Sunshine matrix — including that Moonlight has no end-to-end
number and its "network latency" is an ENet control RTT, so punktfunk's
headline must not be compared against any single Moonlight line.
Verified here: linux client + probe + core check/clippy/fmt green, android
native cargo-ndk arm64 check green. Pending: Windows CI + on-glass, swift
test on the mac, on-device Android.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Settings gains a GPU selector (shown only on multi-GPU boxes): the picked
DXGI adapter drives decode + present, persisted as Settings.adapter and
applied at the next stream - gpu.rs now caches the shared device keyed by
the resolved preference (env PUNKTFUNK_ADAPTER > Settings > the window's
monitor's adapter) so a change needs no app restart.
- Ctrl+Alt+Shift+D disconnects the session (consumed locally, captured or
released): the hook releases capture and trips the session stop flag,
plumbed through the stream-page handoff; the pump winds down and the UI
navigates back to the host list.
- Stream HUD extended: codec chip (HEVC/H.264/AV1), display-side line from
the render thread (presents/s + capture-to-decoded vs capture-to-on-glass
p50), session line (host name, duration, network-lost frames, skipped
backlog frames), and both shortcut hints incl. the new disconnect.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Decode+present rewrite (first real pixels on glass for this client):
- Decode: FFmpeg D3D11VA on NVIDIA/AMD/Intel. get_format now only returns
AV_PIX_FMT_D3D11 and lets libavcodec build the decode pool from
hw_device_ctx (hand-built frames contexts failed three different ways:
NVIDIA rejects DECODER|SHADER_RESOURCE arrays, BindFlags=0 fails texture
creation, Intel rejects non-128-aligned HEVC surfaces at the first
SubmitDecoderBuffers). A DXVA profile probe before the hwdevice commits
hardware-vs-software up front instead of burning the opening IDR;
extra_hw_frames covers the frames the client holds.
- Present: the decoded slice is copied with ONE display-size-boxed
CopySubresourceRegion (a planar slice is a single subresource in D3D11;
the old two-copy D3D12-style code silently no-opped - the black screen)
into a sampleable NV12/P010 texture, per-plane SRVs + YUV->RGB shaders.
- New dedicated render thread (render.rs): presenting is decoupled from the
XAML thread; frame-latency-waitable swapchain + SetMaximumFrameLatency(1),
newest-wins drain after the wait, crossbeam frame channel with pts for a
capture->presented p50 log.
- HiDPI: pixel-sized buffers + SetMatrixTransform(96/dpi) - was blurry at
125/150 % scaling.
- Software fallback now feeds the same shaders (swscale -> NV12/P010 planes
-> two dynamic plane textures); ps_rgba/X2BGR10 path deleted, hw/sw colour
math identical.
- Adapter selection for hybrid boxes: PUNKTFUNK_ADAPTER > the window's
monitor's adapter > default; PUNKTFUNK_D3D_DEBUG=1 debug layer.
- Session pump: request_keyframe at start and on hw->sw demotion (infinite
GOP would otherwise sit on a black screen).
Validated live on the Arc Pro + RTX 3500 Ada laptop against the local
Windows host: 60 fps D3D11VA on both vendors, software path, GUI on glass.
Also: embedded app icon (build.rs winresource + WM_SETICON, MSIX
Square44x44 targetsize assets, pack-msix stages them) and the hosts-page
tile rework (tap-to-connect tiles with sibling overflow menu - fixes
forget-also-connects - in-tile rename editor, add-host modal via root state).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
enricobuehler