feat(client): freeze-until-reanchor loss recovery on Android + Apple via shared core gate
After unrecoverable loss the host keeps sending delta frames that reference a picture the client never received; hardware decoders conceal these as gray/ garbage with a success status. Linux already withheld them and held the last good frame until a proven clean re-anchor — this brings that behavior to the Android and Apple clients. Extract the Linux pump's freeze state machine into a shared `ReanchorGate` in punktfunk-core (reanchor.rs, 18 tests) exposed over the C ABI (ABI v6, additive — no wire change) for the Swift clients. Migrate the Linux/Deck pump (pf-client-core) onto it as the parity proof (no-op refactor). Then wire: - Android (decode.rs, both sync + async loops): arm on the frame-index gap, a pts-keyed flag map carries the wire flags to the output-buffer release, fold the gate per drained output, gate.poll replaces the dropped-climb block. - Apple Stage2Pipeline (default): arm on a gap (new noteFrameIndexGap), withhold at the ring-submit seam (CAMetalLayer holds its last drawable), poll framesDropped, fold VT decode errors through the no-output streak. - Apple StreamPump (stage-1): fold at enqueue, withhold via kCMSampleAttachmentKey_DoNotDisplay so the layer keeps decoding (reference chain intact) but holds the last displayed frame. - Apple VideoDecoder: thread the AU's wire flags to the async decode callback via a retained FrameContext refcon (replaces the receivedNs bit-pattern scalar). Lifts only on a proven re-anchor (IDR / RFI anchor / 2nd recovery mark) with a 500 ms backstop so a lost re-anchor can never freeze forever. Apple: swift build clean, 123/123 tests pass (incl. VideoToolboxRoundTripTests). On-glass loss-injection validation still owed. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
@@ -15,6 +15,7 @@ use ndk::media::media_format::MediaFormat;
|
||||
use ndk::native_window::NativeWindow;
|
||||
use punktfunk_core::client::NativeClient;
|
||||
use punktfunk_core::error::PunktfunkError;
|
||||
use punktfunk_core::reanchor::{GateVerdict, ReanchorGate};
|
||||
use punktfunk_core::session::Frame;
|
||||
use std::collections::VecDeque;
|
||||
use std::ffi::c_void;
|
||||
@@ -208,9 +209,15 @@ fn run_sync(
|
||||
// pressure the AU stays parked here instead of being dropped (a drop forces a keyframe
|
||||
// round-trip) and we only pop the next one once it's queued.
|
||||
let mut pending: Option<Frame> = None;
|
||||
// Loss recovery: watch the host→client unrecoverable-drop count and ask for an IDR when it
|
||||
// climbs.
|
||||
let mut last_dropped = client.frames_dropped();
|
||||
// Freeze-until-reanchor: the shared post-loss gate ([`punktfunk_core::reanchor::ReanchorGate`]).
|
||||
// Armed on a frame-index gap or a dropped-count climb, it withholds the decoder's concealed output
|
||||
// (released WITHOUT rendering — the SurfaceView keeps the last rendered frame on glass) until a
|
||||
// proven clean re-anchor lifts it: an IDR (wire FLAG_SOF), an RFI anchor, or the 2nd recovery mark.
|
||||
// `last_kf_req` throttles the keyframe intents it emits; `recovery_flags` carries each AU's
|
||||
// user_flags from feed to present (keyed by the codec-echoed pts) so `on_decoded` reads the
|
||||
// re-anchor signalling the platform decoder doesn't expose.
|
||||
let mut gate = ReanchorGate::new(client.frames_dropped());
|
||||
let mut recovery_flags: VecDeque<(u64, u32)> = VecDeque::new();
|
||||
let mut last_kf_req: Option<Instant> = None;
|
||||
// Skew-corrected latency stats (spec: design/stats-unification.md) use the negotiated
|
||||
// host-minus-client clock offset (0 if the host didn't answer the skew handshake — then the
|
||||
@@ -245,9 +252,18 @@ fn run_sync(
|
||||
Ok(frame) => {
|
||||
// Loss recovery (RFI): feed the frame index so a forward gap fires a throttled
|
||||
// reference-frame-invalidation request — an RFI-capable host (AMD LTR / NVENC)
|
||||
// recovers with a cheap clean P-frame instead of a full IDR. The frames_dropped
|
||||
// keyframe path below stays the backstop when the recovery frame itself is lost.
|
||||
let _ = client.note_frame_index(frame.frame_index);
|
||||
// recovers with a cheap clean P-frame instead of a full IDR. The same forward gap
|
||||
// arms the freeze gate so the decoder's concealment is held off the screen until the
|
||||
// recovery re-anchors. The frames_dropped keyframe path below stays the backstop.
|
||||
if client.note_frame_index(frame.frame_index) {
|
||||
gate.arm(Instant::now());
|
||||
}
|
||||
// Park this AU's re-anchor flags for the present side (keyed by the pts the codec
|
||||
// echoes on the output buffer) — unconditional, unlike the HUD's `in_flight` map.
|
||||
recovery_flags.push_back((frame.pts_ns / 1000, frame.flags));
|
||||
if recovery_flags.len() > IN_FLIGHT_CAP {
|
||||
recovery_flags.pop_front();
|
||||
}
|
||||
if fed == 0 {
|
||||
let p = &frame.data;
|
||||
log::info!(
|
||||
@@ -336,6 +352,8 @@ fn run_sync(
|
||||
&mut in_flight,
|
||||
clock_offset.load(Ordering::Relaxed),
|
||||
&tracker,
|
||||
&mut gate,
|
||||
&mut recovery_flags,
|
||||
);
|
||||
rendered += r;
|
||||
discarded += d;
|
||||
@@ -375,21 +393,19 @@ fn run_sync(
|
||||
work_accum_ns = 0;
|
||||
}
|
||||
|
||||
// Loss recovery: under infinite GOP the only recovery keyframe is one we request. The
|
||||
// reassembler drops unrecoverable AUs (frames_dropped); the decoder then conceals the
|
||||
// reference-missing delta frames that follow and renders them without error, so keying off
|
||||
// a decode error rarely fires. Request an IDR when the drop count climbs, throttled — the
|
||||
// decode stays wedged for several frames until the IDR lands, so requesting every frame
|
||||
// would flood the control stream.
|
||||
let dropped = client.frames_dropped();
|
||||
if dropped > last_dropped {
|
||||
last_dropped = dropped;
|
||||
// Loss recovery + overdue backstop, folded through the gate. Under infinite GOP the only
|
||||
// recovery keyframe is one we request; the reassembler drops unrecoverable AUs (frames_dropped)
|
||||
// and the decoder then conceals the reference-missing deltas and renders them without error, so
|
||||
// a decode-error trigger rarely fires — the gate arms the freeze on the drop-count climb
|
||||
// instead. An overdue freeze (held REANCHOR_FREEZE_MAX with no clean re-anchor) re-asks while it
|
||||
// keeps holding: never resume to gray — a dead stream is the QUIC idle-timeout watchdog's job.
|
||||
let now = Instant::now();
|
||||
if last_kf_req.is_none_or(|t| now.duration_since(t) >= Duration::from_millis(100)) {
|
||||
if gate.poll(client.frames_dropped(), now)
|
||||
&& last_kf_req.is_none_or(|t| now.duration_since(t) >= Duration::from_millis(100))
|
||||
{
|
||||
last_kf_req = Some(now);
|
||||
let _ = client.request_keyframe();
|
||||
log::debug!("decode: requested keyframe (loss recovery, dropped={dropped})");
|
||||
}
|
||||
log::debug!("decode: requested keyframe (loss recovery / overdue re-anchor)");
|
||||
}
|
||||
}
|
||||
|
||||
@@ -707,8 +723,10 @@ struct OutputReady {
|
||||
/// internal looper thread) push the codec ones; the feeder thread pushes `Au`. Each carries only
|
||||
/// owned/`Copy` data so the callback closures satisfy the `Send` bound and never touch the codec.
|
||||
enum DecodeEvent {
|
||||
/// A received access unit from the feeder, ready to queue into the decoder.
|
||||
Au(Frame),
|
||||
/// A received access unit from the feeder, ready to queue into the decoder. The `bool` is the
|
||||
/// feeder's [`NativeClient::note_frame_index`] verdict — `true` when this AU revealed a forward
|
||||
/// frame-index gap, so the loop arms the freeze gate (the feeder already fired the RFI request).
|
||||
Au(Frame, bool),
|
||||
/// An input buffer slot freed (index) — we can queue an AU into it.
|
||||
InputAvailable(usize),
|
||||
/// A decoded frame is ready (buffer index + echoed pts + the callback-time `decoded` stamp).
|
||||
@@ -894,7 +912,12 @@ fn run_async(
|
||||
let mut discarded: u64 = 0;
|
||||
// AUs larger than the codec input buffer, dropped whole (see `feed`/`feed_ready`).
|
||||
let mut oversized_dropped: u64 = 0;
|
||||
let mut last_dropped = client.frames_dropped();
|
||||
// Freeze-until-reanchor gate (see the sync loop for the rationale). Armed on a frame-index gap
|
||||
// (the feeder's Au verdict), a parked-AU overflow drop, a dropped-count climb, or a recoverable
|
||||
// codec error; `recovery_flags` carries each AU's user_flags from `dispatch_event` (feed) to
|
||||
// `present_ready` (present), keyed by the codec-echoed pts.
|
||||
let mut gate = ReanchorGate::new(client.frames_dropped());
|
||||
let mut recovery_flags: VecDeque<(u64, u32)> = VecDeque::new();
|
||||
let mut last_kf_req: Option<Instant> = None;
|
||||
// Productive (dispatch+feed+present) time between displayed frames; reported to ADPF once one is
|
||||
// presented. The blocking event wait is excluded (idle, not work) — same accounting as the sync loop.
|
||||
@@ -920,6 +943,8 @@ fn run_async(
|
||||
&mut ready,
|
||||
&mut fmt_dirty,
|
||||
&mut fatal,
|
||||
&mut gate,
|
||||
&mut recovery_flags,
|
||||
));
|
||||
}
|
||||
// Coalesce every other event already queued into this one work pass — correct newest-only
|
||||
@@ -932,6 +957,8 @@ fn run_async(
|
||||
&mut ready,
|
||||
&mut fmt_dirty,
|
||||
&mut fatal,
|
||||
&mut gate,
|
||||
&mut recovery_flags,
|
||||
));
|
||||
}
|
||||
stats.note_skipped(aus_dropped); // parked-AU overflow drops are client-side skips too
|
||||
@@ -956,6 +983,8 @@ fn run_async(
|
||||
&tracker,
|
||||
&mut rendered,
|
||||
&mut discarded,
|
||||
&mut gate,
|
||||
&mut recovery_flags,
|
||||
);
|
||||
|
||||
work_accum_ns += work_t0.elapsed().as_nanos() as i64;
|
||||
@@ -987,19 +1016,21 @@ fn run_async(
|
||||
log::info!("decode: fed={fed} rendered={rendered} discarded={discarded}");
|
||||
}
|
||||
}
|
||||
// Loss recovery: request an IDR when the reassembler's unrecoverable-drop count climbs (or we
|
||||
// dropped a parked AU on overflow), throttled so a multi-frame recovery gap doesn't flood the
|
||||
// control stream.
|
||||
let dropped = client.frames_dropped();
|
||||
if dropped > last_dropped || aus_dropped > 0 {
|
||||
last_dropped = dropped;
|
||||
// Loss recovery + overdue backstop, folded through the gate. A parked-AU overflow drop is itself
|
||||
// a loss, so it arms the freeze directly; the gate's `poll` then arms on a dropped-count climb
|
||||
// and re-asks on an overdue freeze. All keyframe intents route through the shared 100 ms
|
||||
// throttle so a multi-frame recovery gap can't flood the control stream.
|
||||
let now = Instant::now();
|
||||
if last_kf_req.is_none_or(|t| now.duration_since(t) >= Duration::from_millis(100)) {
|
||||
if aus_dropped > 0 {
|
||||
gate.arm(now);
|
||||
}
|
||||
if (gate.poll(client.frames_dropped(), now) || aus_dropped > 0)
|
||||
&& last_kf_req.is_none_or(|t| now.duration_since(t) >= Duration::from_millis(100))
|
||||
{
|
||||
last_kf_req = Some(now);
|
||||
let _ = client.request_keyframe();
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
let _ = codec.stop();
|
||||
shutdown.store(true, Ordering::SeqCst); // ensure the feeder wakes and exits, then join it
|
||||
@@ -1033,8 +1064,9 @@ fn feeder_loop(
|
||||
Ok(frame) => {
|
||||
// Loss recovery (RFI): a forward frame-index gap fires a throttled reference-frame-
|
||||
// invalidation request so an RFI-capable host recovers with a cheap clean P-frame
|
||||
// instead of a full IDR (the frames_dropped keyframe path is the backstop).
|
||||
let _ = client.note_frame_index(frame.frame_index);
|
||||
// instead of a full IDR (the frames_dropped keyframe path is the backstop). The gap
|
||||
// verdict rides the Au event so the decode loop arms its freeze gate on the same signal.
|
||||
let gap = client.note_frame_index(frame.frame_index);
|
||||
if stats.enabled() {
|
||||
let received_ns = now_realtime_ns();
|
||||
let clock_offset = clock_offset.load(Ordering::Relaxed) as i128;
|
||||
@@ -1067,7 +1099,7 @@ fn feeder_loop(
|
||||
}
|
||||
}
|
||||
}
|
||||
if ev_tx.send(DecodeEvent::Au(frame)).is_err() {
|
||||
if ev_tx.send(DecodeEvent::Au(frame, gap)).is_err() {
|
||||
break; // the decode loop is gone
|
||||
}
|
||||
}
|
||||
@@ -1079,6 +1111,7 @@ fn feeder_loop(
|
||||
|
||||
/// Route one [`DecodeEvent`] into the loop's working sets. Returns `true` only when a parked AU was
|
||||
/// dropped on overflow (the caller then requests a keyframe).
|
||||
#[allow(clippy::too_many_arguments)] // two call sites; the freeze gate + flag map are threaded in
|
||||
fn dispatch_event(
|
||||
ev: DecodeEvent,
|
||||
pending_aus: &mut VecDeque<Frame>,
|
||||
@@ -1086,9 +1119,20 @@ fn dispatch_event(
|
||||
ready: &mut Vec<OutputReady>,
|
||||
fmt_dirty: &mut bool,
|
||||
fatal: &mut bool,
|
||||
gate: &mut ReanchorGate,
|
||||
recovery_flags: &mut VecDeque<(u64, u32)>,
|
||||
) -> bool {
|
||||
match ev {
|
||||
DecodeEvent::Au(f) => {
|
||||
DecodeEvent::Au(f, gap) => {
|
||||
// A forward frame-index gap arms the freeze; park this AU's flags for the present side to
|
||||
// fold `on_decoded` (keyed by the pts the codec will echo).
|
||||
if gap {
|
||||
gate.arm(Instant::now());
|
||||
}
|
||||
recovery_flags.push_back((f.pts_ns / 1000, f.flags));
|
||||
if recovery_flags.len() > IN_FLIGHT_CAP {
|
||||
recovery_flags.pop_front();
|
||||
}
|
||||
pending_aus.push_back(f);
|
||||
if pending_aus.len() > FRAME_PARK_CAP {
|
||||
pending_aus.pop_front(); // sustained overflow — drop oldest, signal a keyframe request
|
||||
@@ -1109,6 +1153,10 @@ fn dispatch_event(
|
||||
DecodeEvent::Error { fatal: f } => {
|
||||
if f {
|
||||
*fatal = true;
|
||||
} else {
|
||||
// A recoverable/transient codec error is a decode hiccup on a broken reference chain —
|
||||
// arm the freeze so the concealed output it recovers into is held off the screen.
|
||||
gate.arm(Instant::now());
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -1180,6 +1228,8 @@ fn present_ready(
|
||||
tracker: &DisplayTracker,
|
||||
rendered: &mut u64,
|
||||
discarded: &mut u64,
|
||||
gate: &mut ReanchorGate,
|
||||
recovery_flags: &mut VecDeque<(u64, u32)>,
|
||||
) {
|
||||
if ready.is_empty() {
|
||||
return;
|
||||
@@ -1192,10 +1242,16 @@ fn present_ready(
|
||||
note_decoded_pts(stats, &mut g, clock_offset, o.pts_us, o.decoded_ns);
|
||||
}
|
||||
}
|
||||
// Fold EVERY output through the gate in pts (== decode) order — even the ones newest-wins discards —
|
||||
// so the two-mark re-anchor count stays correct; the newest's verdict decides whether it reaches
|
||||
// glass (`false` = withheld concealment; the SurfaceView keeps the last rendered frame frozen on).
|
||||
let now = Instant::now();
|
||||
let last = ready.len() - 1;
|
||||
let mut skipped: u64 = 0;
|
||||
for (i, o) in ready.drain(..).enumerate() {
|
||||
let render = i == last;
|
||||
let flags = take_flags(recovery_flags, o.pts_us);
|
||||
let present = gate.on_decoded(flags, false, now) == GateVerdict::Present;
|
||||
let render = i == last && present;
|
||||
match codec.release_output_buffer_by_index(o.index, render) {
|
||||
Ok(()) if render => {
|
||||
*rendered += 1;
|
||||
@@ -1215,7 +1271,7 @@ fn present_ready(
|
||||
}
|
||||
}
|
||||
}
|
||||
stats.note_skipped(skipped); // HUD `skipped` counter (newest-wins drops); no-op while hidden
|
||||
stats.note_skipped(skipped); // HUD `skipped` counter (newest-wins + held-off drops); no-op hidden
|
||||
}
|
||||
|
||||
/// React to an output-format change by signalling the stream's HDR dataspace on the Surface (SDR
|
||||
@@ -1411,19 +1467,28 @@ fn drain(
|
||||
in_flight: &mut VecDeque<(u64, i128)>,
|
||||
clock_offset: i64,
|
||||
tracker: &DisplayTracker,
|
||||
gate: &mut ReanchorGate,
|
||||
recovery_flags: &mut VecDeque<(u64, u32)>,
|
||||
) -> (u64, u64) {
|
||||
// Newest ready buffer so far (presented after the loop) with its HUD metadata —
|
||||
// `Some((pts_us, decoded_ns))` only while the HUD is visible (the stamp read is gated).
|
||||
// `Some((pts_us, decoded_ns))` only while the HUD is visible. `held_present` is the freeze gate's
|
||||
// verdict for that newest buffer (`false` = a post-loss concealment to withhold).
|
||||
let mut held: Option<(OutputBuffer<'_>, Option<(u64, i128)>)> = None;
|
||||
let mut held_present = true;
|
||||
let mut discarded: u64 = 0;
|
||||
let mut wait = first_wait;
|
||||
loop {
|
||||
match codec.dequeue_output_buffer(wait) {
|
||||
Ok(DequeuedOutputBufferInfoResult::Buffer(buf)) => {
|
||||
wait = Duration::ZERO; // only the first dequeue may block
|
||||
// Fold every dequeued frame through the gate in pts (== decode) order — even the ones
|
||||
// the newest-wins policy discards — so the two-mark re-anchor count stays correct; the
|
||||
// verdict of the newest (last folded) buffer decides whether it reaches glass.
|
||||
let pts_us = buf.info().presentation_time_us().max(0) as u64;
|
||||
let flags = take_flags(recovery_flags, pts_us);
|
||||
held_present = gate.on_decoded(flags, false, Instant::now()) == GateVerdict::Present;
|
||||
let meta = if stats.enabled() {
|
||||
// The dequeue IS the sync loop's decoded-availability instant.
|
||||
let pts_us = buf.info().presentation_time_us().max(0) as u64;
|
||||
let decoded_ns = now_realtime_ns();
|
||||
note_decoded_pts(stats, in_flight, clock_offset, pts_us, decoded_ns);
|
||||
Some((pts_us, decoded_ns))
|
||||
@@ -1469,16 +1534,19 @@ fn drain(
|
||||
}
|
||||
}
|
||||
}
|
||||
// Present the newest ready frame, if any, and park its metadata for the render callback.
|
||||
// Present the newest ready frame — UNLESS the gate is withholding it as a post-loss concealment,
|
||||
// in which case release it without rendering (the SurfaceView keeps the last rendered frame frozen
|
||||
// on glass) and count it as a discard rather than a display.
|
||||
let mut rendered = 0;
|
||||
if let Some((buf, meta)) = held {
|
||||
match codec.release_output_buffer(buf, true) {
|
||||
Ok(()) => {
|
||||
match codec.release_output_buffer(buf, held_present) {
|
||||
Ok(()) if held_present => {
|
||||
rendered = 1;
|
||||
if let Some((pts_us, decoded_ns)) = meta {
|
||||
tracker.note_rendered(pts_us, decoded_ns);
|
||||
}
|
||||
}
|
||||
Ok(()) => discarded += 1, // held off the screen — awaiting a clean re-anchor
|
||||
Err(e) => log::warn!("decode: release_output_buffer: {e}"),
|
||||
}
|
||||
}
|
||||
@@ -1520,6 +1588,25 @@ fn note_decoded_pts(
|
||||
stats.note_decoded(e2e_us, decode_us);
|
||||
}
|
||||
|
||||
/// The AU `user_flags` for a decoded output, keyed by the echoed `presentationTimeUs`. Recovery
|
||||
/// signalling (FLAG_SOF IDR marker / RECOVERY_ANCHOR / RECOVERY_POINT) rides the AU's flags, which are
|
||||
/// only in scope at feed time — so the feed side parks `(pts_us, flags)` here and the present side
|
||||
/// looks them up to fold [`ReanchorGate::on_decoded`]. Decode order == input order (low-latency, no
|
||||
/// B-frames), so this evicts entries older than `pts_us` as it goes; a miss (probe filler, or an entry
|
||||
/// aged past the cap) reads `0` — no recovery flags, decoded normally.
|
||||
fn take_flags(map: &mut VecDeque<(u64, u32)>, pts_us: u64) -> u32 {
|
||||
while let Some(&(p, f)) = map.front() {
|
||||
if p > pts_us {
|
||||
break; // future frame — leave it for its own output buffer
|
||||
}
|
||||
map.pop_front();
|
||||
if p == pts_us {
|
||||
return f;
|
||||
}
|
||||
}
|
||||
0
|
||||
}
|
||||
|
||||
/// Map the decoder's reported output colour to a BT.2020 HDR dataspace, or `None` for SDR. The
|
||||
/// integer values are the Android MediaFormat colour constants the NDK shares: COLOR_TRANSFER
|
||||
/// ST2084 = 6 (PQ/HDR10), HLG = 7; COLOR_RANGE FULL = 1, LIMITED = 2 (the host encodes limited).
|
||||
|
||||
@@ -450,6 +450,21 @@ public final class PunktfunkConnection {
|
||||
_ = punktfunk_connection_note_frame_index(h, frameIndex, nil)
|
||||
}
|
||||
|
||||
/// Like `noteFrameIndex`, but also reports whether the core saw a FORWARD frame-index gap — the
|
||||
/// signal that intervening frames were lost and the following AUs reference a picture that never
|
||||
/// arrived. The post-loss re-anchor gate arms its display freeze on a gap (the earliest, most
|
||||
/// precise loss trigger — ahead of the `framesDropped` climb). Same core side effect as
|
||||
/// `noteFrameIndex` (the throttled RFI request); call it for every received AU. Returns false
|
||||
/// after close.
|
||||
public func noteFrameIndexGap(_ frameIndex: UInt32) -> Bool {
|
||||
abiLock.lock()
|
||||
defer { abiLock.unlock() }
|
||||
guard let h = handle, !closeRequested else { return false }
|
||||
var gap = false
|
||||
_ = punktfunk_connection_note_frame_index(h, frameIndex, &gap)
|
||||
return gap
|
||||
}
|
||||
|
||||
/// Cumulative access units the host→client reassembler dropped as unrecoverable (FEC couldn't
|
||||
/// rebuild them). The video pump polls this and calls `requestKeyframe()` when it climbs — the
|
||||
/// correct loss trigger under the host's infinite GOP, where unrecoverable loss yields
|
||||
|
||||
@@ -0,0 +1,99 @@
|
||||
// Swift wrapper around the punktfunk-core C ABI's post-loss re-anchor gate
|
||||
// (`punktfunk_reanchor_gate_*`, ABI v6). The shared Rust gate (crates/punktfunk-core/src/reanchor.rs)
|
||||
// is what the Linux/Windows desktop pump and the Android client use directly; the Swift clients reach
|
||||
// it across the C ABI so the freeze-until-reanchor policy is defined ONCE for every platform.
|
||||
//
|
||||
// Why a freeze at all: after unrecoverable loss the host keeps sending delta frames that reference a
|
||||
// picture the client never got. Hardware decoders (VideoToolbox included) don't reliably error on
|
||||
// that — they CONCEAL, returning a gray/garbage frame with a success status. Presenting those is the
|
||||
// visible "gray flash with motion" of the loss reports. The gate withholds concealed frames and holds
|
||||
// the last good picture on glass until a PROVEN clean re-anchor lands — an IDR (wire `FLAG_SOF`), an
|
||||
// RFI recovery anchor (`USER_FLAG_RECOVERY_ANCHOR`), or the 2nd of two intra-refresh recovery marks
|
||||
// (`USER_FLAG_RECOVERY_POINT`) — with a bounded backstop so a lost re-anchor can never freeze forever.
|
||||
// See punktfunk-planning design/client-reanchor-freeze-parity.md.
|
||||
//
|
||||
// Threading: one gate per session. Its calls arrive from two threads — the pump thread (`arm` on a
|
||||
// frame-index gap / a submit failure, `poll` per iteration) and a VideoToolbox decode thread
|
||||
// (`onDecoded` per decoded frame, `onNoOutput` on a decode error). The raw Rust gate is a plain
|
||||
// struct behind an opaque pointer with no internal synchronization, so every call is serialized under
|
||||
// `lock` here — the calls are cheap field updates, so contention is negligible. `@unchecked Sendable`:
|
||||
// the lock enforces the contract.
|
||||
|
||||
import Foundation
|
||||
import PunktfunkCore
|
||||
|
||||
final class ReanchorGate: @unchecked Sendable {
|
||||
private let lock = NSLock()
|
||||
/// The opaque `ReanchorGate *`. `var` so `reseed` can swap it at session start. Never NULL
|
||||
/// (`punktfunk_reanchor_gate_new` never returns NULL).
|
||||
private var ptr: OpaquePointer
|
||||
|
||||
/// Seed the baseline with the connection's current `framesDropped` so the first `poll` doesn't
|
||||
/// read the session's starting drop count as a fresh loss.
|
||||
init(framesDropped: UInt64) {
|
||||
ptr = punktfunk_reanchor_gate_new(framesDropped)
|
||||
}
|
||||
|
||||
deinit { punktfunk_reanchor_gate_free(ptr) }
|
||||
|
||||
/// Re-anchor the drop-count baseline to `framesDropped` for a (re)started session. The gate is
|
||||
/// created in the pipeline's init (before a connection exists, seeded 0); `start` calls this once
|
||||
/// the live connection's count is known so a mid-life connection's non-zero baseline isn't
|
||||
/// mistaken for loss on the first poll.
|
||||
func reseed(framesDropped: UInt64) {
|
||||
lock.lock()
|
||||
defer { lock.unlock() }
|
||||
punktfunk_reanchor_gate_free(ptr)
|
||||
ptr = punktfunk_reanchor_gate_new(framesDropped)
|
||||
}
|
||||
|
||||
/// Arm the freeze: a loss was detected (a frame-index gap, or a decoder wedge). Zeroes the
|
||||
/// recovery-mark count and (re)sets the backstop deadline.
|
||||
func arm() {
|
||||
lock.lock()
|
||||
punktfunk_reanchor_gate_arm(ptr)
|
||||
lock.unlock()
|
||||
}
|
||||
|
||||
/// Fold one decoded frame. `flags` is the AU's wire `user_flags`. Returns true to PRESENT the
|
||||
/// frame, false to WITHHOLD it as a post-loss concealment (hold the last good picture). Pass
|
||||
/// `decoderKeyframe: false` — VideoToolbox doesn't flag IDRs, so the wire `FLAG_SOF` covers it.
|
||||
func onDecoded(flags: UInt32, decoderKeyframe: Bool = false) -> Bool {
|
||||
lock.lock()
|
||||
defer { lock.unlock() }
|
||||
var present = false
|
||||
_ = punktfunk_reanchor_gate_on_decoded(ptr, flags, decoderKeyframe, &present)
|
||||
return present
|
||||
}
|
||||
|
||||
/// A received AU produced no decoded frame (a VideoToolbox decode error). Returns true when the
|
||||
/// no-output streak has tripped (the gate armed the freeze) and the caller should — throttled —
|
||||
/// request a keyframe.
|
||||
func onNoOutput() -> Bool {
|
||||
lock.lock()
|
||||
defer { lock.unlock() }
|
||||
var requestKf = false
|
||||
_ = punktfunk_reanchor_gate_on_no_output(ptr, &requestKf)
|
||||
return requestKf
|
||||
}
|
||||
|
||||
/// Periodic fold of the session's `framesDropped` plus the overdue backstop. Returns true when the
|
||||
/// caller should — throttled — request a keyframe (a drop-count climb armed a fresh freeze, or the
|
||||
/// freeze is overdue and re-asks while it keeps holding).
|
||||
func poll(framesDropped: UInt64) -> Bool {
|
||||
lock.lock()
|
||||
defer { lock.unlock() }
|
||||
var requestKf = false
|
||||
_ = punktfunk_reanchor_gate_poll(ptr, framesDropped, &requestKf)
|
||||
return requestKf
|
||||
}
|
||||
|
||||
/// Whether the gate is currently withholding concealed frames (frozen on the last good picture).
|
||||
var isHolding: Bool {
|
||||
lock.lock()
|
||||
defer { lock.unlock() }
|
||||
var holding = false
|
||||
_ = punktfunk_reanchor_gate_is_holding(ptr, &holding)
|
||||
return holding
|
||||
}
|
||||
}
|
||||
@@ -259,6 +259,10 @@ public final class Stage2Pipeline {
|
||||
private let endToEndMeter: LatencyMeter?
|
||||
private let displayMeter: LatencyMeter?
|
||||
private let recovery = KeyframeRecovery()
|
||||
/// Post-loss freeze-until-reanchor gate (shared core policy via the C ABI). Created here seeded 0;
|
||||
/// `start` reseeds it to the live connection's drop count. Captured by the decoder callbacks
|
||||
/// (which withhold concealed frames) and driven by the pump (arm on a gap, poll per iteration).
|
||||
private let gate = ReanchorGate(framesDropped: 0)
|
||||
private var token = StopFlag()
|
||||
private var offsetNs: Int64 = 0
|
||||
/// Signalled when the pump thread exits, so `stop()` can join it (bounded) before `decoder.reset()`
|
||||
@@ -306,21 +310,29 @@ public final class Stage2Pipeline {
|
||||
let ring = ring
|
||||
let recovery = recovery
|
||||
let renderSignal = renderSignal
|
||||
let gate = gate
|
||||
self.decoder = VideoDecoder(
|
||||
onDecoded: { frame in
|
||||
// Decode stage = received→decoded, both client CLOCK_REALTIME (offset 0 — no
|
||||
// skew applies). Stamped at decode completion, so it covers every decoded frame,
|
||||
// including ones the newest-wins ring drops before present.
|
||||
// including ones the re-anchor gate withholds or the newest-wins ring drops.
|
||||
decodeMeter?.record(
|
||||
ptsNs: UInt64(frame.receivedNs), atNs: frame.decodedNs, offsetNs: 0)
|
||||
// Freeze-until-reanchor: WITHHOLD a decoder-concealed post-loss frame (the gray/
|
||||
// garbage VideoToolbox returns Ok for a reference-missing delta) — don't submit it,
|
||||
// so the CAMetalLayer keeps its last good drawable on glass. The gate lifts (returns
|
||||
// present) on a proven clean re-anchor (IDR / RFI anchor / 2nd recovery mark) or the
|
||||
// bounded backstop. decoderKeyframe=false: VT doesn't flag IDRs, the wire FLAG_SOF does.
|
||||
guard gate.onDecoded(flags: frame.flags) else { return }
|
||||
ring.submit(frame)
|
||||
// FRAME ARRIVAL is the render trigger (never the display link — see the header).
|
||||
renderSignal.signal()
|
||||
},
|
||||
// Async decode failure (a bad P-frame referencing a lost/corrupt IDR): the pump resets to
|
||||
// re-gate on the next IDR, and we ask the host to send one now (infinite GOP — it wouldn't
|
||||
// Async decode failure (a bad P-frame referencing a lost/corrupt IDR): fold it into the
|
||||
// gate's no-output streak (which arms the freeze after a short run, matching the desktop),
|
||||
// and when that trips ask the host for a fresh IDR now (infinite GOP — it wouldn't
|
||||
// otherwise come soon). Throttled in KeyframeRecovery.
|
||||
onDecodeError: { _ in recovery.request() })
|
||||
onDecodeError: { _ in if gate.onNoOutput() { recovery.request() } })
|
||||
}
|
||||
|
||||
/// Start pulling AUs into the decoder. MAIN THREAD. `onFrame` fires per AU at receipt (the
|
||||
@@ -334,6 +346,7 @@ public final class Stage2Pipeline {
|
||||
) {
|
||||
offsetNs = connection.clockOffsetNs
|
||||
recovery.bind(connection) // arm host-keyframe recovery for this session
|
||||
gate.reseed(framesDropped: connection.framesDropped()) // baseline the freeze to this session
|
||||
token = StopFlag() // fresh token per start — a stop is permanent (like StreamPump)
|
||||
|
||||
// Configure the decoder's chroma + the layer's initial colorimetry before the first frame. The
|
||||
@@ -348,6 +361,7 @@ public final class Stage2Pipeline {
|
||||
let recovery = recovery
|
||||
let presenter = presenter
|
||||
let pumpStopped = pumpStopped
|
||||
let reanchorGate = gate
|
||||
let thread = Thread {
|
||||
defer { pumpStopped.signal() } // let stop() join the pump (bounded) before decoder.reset()
|
||||
var format: CMVideoFormatDescription?
|
||||
@@ -379,6 +393,9 @@ public final class Stage2Pipeline {
|
||||
awaitingIDR = true
|
||||
}
|
||||
if awaitingIDR { recovery.request() }
|
||||
// Freeze backstop: a drop-count climb arms the gate (in case the frame-index gap
|
||||
// below was itself lost), and an overdue freeze re-asks for the re-anchor.
|
||||
if reanchorGate.poll(framesDropped: dropped) { recovery.request() }
|
||||
// Drain HDR mastering metadata (0xCE) and hand it to the PRESENTER (→ CAEDRMetadata).
|
||||
// Polled UNCONDITIONALLY (not gated on connection.isHDR, the fixed Welcome flag): the
|
||||
// host sends 0xCE only for HDR, INCLUDING a mid-session SDR→HDR transition (a game
|
||||
@@ -391,8 +408,10 @@ public final class Stage2Pipeline {
|
||||
// Loss recovery (RFI): a forward frame-index gap fires a throttled reference-
|
||||
// frame-invalidation request so an RFI-capable host (AMD LTR / NVENC) recovers
|
||||
// with a cheap clean P-frame instead of a full IDR. The framesDropped-driven
|
||||
// recovery below stays the backstop for when the recovery frame itself is lost.
|
||||
connection.noteFrameIndex(au.frameIndex)
|
||||
// recovery above stays the backstop for when the recovery frame itself is lost.
|
||||
// The same gap is the earliest, most precise signal to ARM the display freeze —
|
||||
// the following concealed frames are withheld until a clean re-anchor.
|
||||
if connection.noteFrameIndexGap(au.frameIndex) { reanchorGate.arm() }
|
||||
onFrame?(au)
|
||||
if let f = connection.videoCodec.formatDescription(fromKeyframe: au.data) {
|
||||
format = f // refreshed on every IDR (mode changes included)
|
||||
|
||||
@@ -28,6 +28,11 @@ final class StreamPump {
|
||||
// Coalesced host keyframe requests (100 ms throttle — see KeyframeRecovery).
|
||||
let recovery = KeyframeRecovery()
|
||||
recovery.bind(connection)
|
||||
// Post-loss freeze-until-reanchor (shared core policy via the C ABI). Stage-1 has no per-frame
|
||||
// decode callback, so the gate is folded at ENQUEUE (from the AU's wire flags): a withheld
|
||||
// frame is still enqueued but flagged DoNotDisplay so the layer's decoder keeps the reference
|
||||
// chain fed while the last GOOD picture stays on glass — until a clean re-anchor lifts it.
|
||||
let gate = ReanchorGate(framesDropped: connection.framesDropped())
|
||||
// The layer is non-Sendable but its enqueue/flush are documented thread-safe, and after
|
||||
// this point only the pump thread drives it — assert that so the @Sendable Thread closure
|
||||
// may capture it.
|
||||
@@ -77,13 +82,17 @@ final class StreamPump {
|
||||
awaitingIDR = true
|
||||
}
|
||||
if awaitingIDR { recovery.request() }
|
||||
// Freeze backstop: a drop-count climb arms the gate (should the frame-index gap
|
||||
// below be lost too), and an overdue freeze re-asks for the re-anchor.
|
||||
if gate.poll(framesDropped: dropped) { recovery.request() }
|
||||
|
||||
guard let au = try connection.nextAU(timeoutMs: 100) else { return true }
|
||||
// Loss recovery (RFI): a forward frame-index gap fires a throttled reference-
|
||||
// frame-invalidation request so an RFI-capable host (AMD LTR / NVENC) recovers
|
||||
// with a cheap clean P-frame instead of a full IDR. The framesDropped-driven
|
||||
// recovery above stays the backstop for when the recovery frame itself is lost.
|
||||
connection.noteFrameIndex(au.frameIndex)
|
||||
// The same gap is the earliest, most precise signal to ARM the display freeze.
|
||||
if connection.noteFrameIndexGap(au.frameIndex) { gate.arm() }
|
||||
onFrame?(au)
|
||||
let idrFormat = connection.videoCodec.formatDescription(fromKeyframe: au.data)
|
||||
if let f = idrFormat {
|
||||
@@ -107,6 +116,7 @@ final class StreamPump {
|
||||
// delta into a failed layer can't recover it.
|
||||
if !wasFailed { pumpLog.warning("video: display layer .failed — flushing + re-anchoring") }
|
||||
layer.flush()
|
||||
gate.arm() // a wedged decoder is a loss — freeze until the re-anchor
|
||||
if idrFormat == nil {
|
||||
format = nil
|
||||
awaitingIDR = true
|
||||
@@ -117,6 +127,13 @@ final class StreamPump {
|
||||
let sample = connection.videoCodec.sampleBuffer(au: au, format: f),
|
||||
!token.isStopped // don't enqueue a stale frame after a restart
|
||||
else { return true }
|
||||
// Freeze-until-reanchor: while holding, WITHHOLD this concealed post-loss frame by
|
||||
// flagging it DoNotDisplay — the layer still decodes it (keeping the reference
|
||||
// chain fed) but shows the last GOOD picture until a clean re-anchor lifts the
|
||||
// gate. Folded from the AU's wire flags (stage-1 has no decode callback).
|
||||
if !gate.onDecoded(flags: au.flags) {
|
||||
StreamPump.setDoNotDisplay(sample)
|
||||
}
|
||||
layer.enqueue(sample)
|
||||
return true
|
||||
} catch {
|
||||
@@ -133,6 +150,21 @@ final class StreamPump {
|
||||
thread.start()
|
||||
}
|
||||
|
||||
/// Flag a sample decode-but-don't-display (`kCMSampleAttachmentKey_DoNotDisplay`). Used to
|
||||
/// withhold decoder-concealed post-loss frames while the re-anchor gate holds: the layer keeps
|
||||
/// its reference chain fed without flipping the frozen picture. No-op if the attachments array
|
||||
/// can't be materialized (then the frame just displays — the freeze degrades to the old behavior).
|
||||
private static func setDoNotDisplay(_ sample: CMSampleBuffer) {
|
||||
guard let attachments = CMSampleBufferGetSampleAttachmentsArray(
|
||||
sample, createIfNecessary: true), CFArrayGetCount(attachments) > 0
|
||||
else { return }
|
||||
let dict = unsafeBitCast(CFArrayGetValueAtIndex(attachments, 0), to: CFMutableDictionary.self)
|
||||
CFDictionarySetValue(
|
||||
dict,
|
||||
Unmanaged.passUnretained(kCMSampleAttachmentKey_DoNotDisplay).toOpaque(),
|
||||
Unmanaged.passUnretained(kCFBooleanTrue).toOpaque())
|
||||
}
|
||||
|
||||
/// Stop pumping (≤ one poll timeout). Does not close the connection.
|
||||
func stop() {
|
||||
token.stop()
|
||||
|
||||
@@ -27,19 +27,40 @@ public struct ReadyFrame: @unchecked Sendable {
|
||||
/// True when the stream is HDR (BT.2020 PQ): the buffer is 10-bit P010 and the presenter must
|
||||
/// configure EDR + BT.2020 PQ output. Derived from the decoded buffer's pixel format.
|
||||
public let isHDR: Bool
|
||||
/// The AU's wire `user_flags` (`AccessUnit.flags`), threaded through the decode via the frame
|
||||
/// context so the re-anchor gate can classify this decoded frame (IDR / RFI anchor / recovery
|
||||
/// mark) at present time — the async decode callback has no other access to it. 0 when unknown.
|
||||
public let flags: UInt32
|
||||
}
|
||||
|
||||
/// Per-frame context threaded through the VideoToolbox frame refcon: the AU's receipt instant (for
|
||||
/// the decode-stage meter) and its wire `user_flags` (for the re-anchor gate). Retained across the
|
||||
/// async decode and reclaimed exactly once — by the output callback for every frame VideoToolbox
|
||||
/// accepts, or by `decode`'s error branch for a frame `DecodeFrame` rejected outright (the callback
|
||||
/// then never fires). A tiny per-frame allocation, the price of smuggling two values (a 64-bit
|
||||
/// instant plus the flags) through the single `void*` a bit-pattern scalar can't hold.
|
||||
private final class FrameContext {
|
||||
let receivedNs: Int64
|
||||
let flags: UInt32
|
||||
init(receivedNs: Int64, flags: UInt32) {
|
||||
self.receivedNs = receivedNs
|
||||
self.flags = flags
|
||||
}
|
||||
}
|
||||
|
||||
/// The C output callback can't capture context, so VideoToolbox hands it the refcon we set at
|
||||
/// session creation — a pointer back to the owning `VideoDecoder`. The per-frame refcon carries
|
||||
/// the AU's `receivedNs` as a pointer bit pattern (a scalar smuggled through the C void*, never
|
||||
/// dereferenced) so the decode stage can be computed against decode-completion.
|
||||
/// session creation — a pointer back to the owning `VideoDecoder`. The per-frame refcon is the
|
||||
/// retained `FrameContext` set at submit; reclaim it here (balancing `passRetained`) and unpack the
|
||||
/// AU's receipt instant (for the decode stage) and wire flags (for the re-anchor gate).
|
||||
private let decoderOutputCallback: VTDecompressionOutputCallback = {
|
||||
refcon, frameRefcon, status, _, imageBuffer, pts, _ in
|
||||
guard let refcon else { return }
|
||||
let receivedNs = frameRefcon.map { Int64(Int(bitPattern: $0)) } ?? 0
|
||||
let ctx = frameRefcon.map { Unmanaged<FrameContext>.fromOpaque($0).takeRetainedValue() }
|
||||
Unmanaged<VideoDecoder>.fromOpaque(refcon)
|
||||
.takeUnretainedValue()
|
||||
.handleDecoded(status: status, imageBuffer: imageBuffer, pts: pts, receivedNs: receivedNs)
|
||||
.handleDecoded(
|
||||
status: status, imageBuffer: imageBuffer, pts: pts,
|
||||
receivedNs: ctx?.receivedNs ?? 0, flags: ctx?.flags ?? 0)
|
||||
}
|
||||
|
||||
/// Owns a `VTDecompressionSession` rebuilt whenever the format description changes (every IDR /
|
||||
@@ -117,16 +138,21 @@ public final class VideoDecoder: @unchecked Sendable {
|
||||
let sample = codec.sampleBuffer(au: au, format: newFormat)
|
||||
else { lock.unlock(); return false }
|
||||
var infoOut = VTDecodeInfoFlags()
|
||||
// The AU's receipt instant + wire flags ride through as a retained context; the output
|
||||
// callback reclaims it. Retain immediately before submit so no early return can leak it.
|
||||
let ctx = FrameContext(receivedNs: au.receivedNs, flags: au.flags)
|
||||
let refcon = Unmanaged.passRetained(ctx).toOpaque()
|
||||
let status = VTDecompressionSessionDecodeFrame(
|
||||
session,
|
||||
sampleBuffer: sample,
|
||||
flags: [._EnableAsynchronousDecompression],
|
||||
// The AU's receipt instant rides through as a bit pattern (nil for 0 — the output
|
||||
// callback maps that back to 0); the callback needs it to stamp the decode stage.
|
||||
frameRefcon: UnsafeMutableRawPointer(bitPattern: Int(au.receivedNs)),
|
||||
frameRefcon: refcon,
|
||||
infoFlagsOut: &infoOut)
|
||||
lock.unlock()
|
||||
if status != noErr {
|
||||
// DecodeFrame rejected the frame outright — the output callback will NOT fire, so
|
||||
// reclaim the context here (balancing passRetained) to avoid leaking it.
|
||||
Unmanaged<FrameContext>.fromOpaque(refcon).release()
|
||||
onDecodeError(status)
|
||||
return false
|
||||
}
|
||||
@@ -231,9 +257,10 @@ public final class VideoDecoder: @unchecked Sendable {
|
||||
}
|
||||
|
||||
/// VT thread. Stamp decode-completion and enqueue, or report the error. `receivedNs` is the
|
||||
/// AU's receipt instant threaded through the frame refcon (0 = unknown).
|
||||
/// AU's receipt instant and `flags` its wire `user_flags`, both threaded through the frame refcon
|
||||
/// (0 = unknown).
|
||||
fileprivate func handleDecoded(
|
||||
status: OSStatus, imageBuffer: CVImageBuffer?, pts: CMTime, receivedNs: Int64
|
||||
status: OSStatus, imageBuffer: CVImageBuffer?, pts: CMTime, receivedNs: Int64, flags: UInt32
|
||||
) {
|
||||
guard status == noErr, let imageBuffer else {
|
||||
onDecodeError(status)
|
||||
@@ -259,6 +286,6 @@ public final class VideoDecoder: @unchecked Sendable {
|
||||
onDecoded(
|
||||
ReadyFrame(
|
||||
ptsNs: ptsNs, receivedNs: receivedNs, decodedNs: decodedNs,
|
||||
pixelBuffer: imageBuffer, isHDR: isHDR))
|
||||
pixelBuffer: imageBuffer, isHDR: isHDR, flags: flags))
|
||||
}
|
||||
}
|
||||
|
||||
@@ -10,6 +10,7 @@ use crate::audio;
|
||||
use crate::video::{DecodedFrame, DecodedImage, Decoder};
|
||||
use punktfunk_core::client::NativeClient;
|
||||
use punktfunk_core::config::{CompositorPref, GamepadPref, Mode};
|
||||
use punktfunk_core::reanchor::{index_gap, GateVerdict, ReanchorGate};
|
||||
use punktfunk_core::PunktfunkError;
|
||||
use std::sync::atomic::{AtomicBool, Ordering};
|
||||
use std::sync::Arc;
|
||||
@@ -99,86 +100,6 @@ pub struct Stats {
|
||||
pub decoder: &'static str,
|
||||
}
|
||||
|
||||
/// Consecutive no-output AUs that force a keyframe request. ~50 ms at 60 Hz — long
|
||||
/// enough not to fire on a one-frame decoder hiccup, short enough that a lost initial
|
||||
/// IDR (or a mid-GOP join) unfreezes almost immediately instead of never.
|
||||
const NO_OUTPUT_KEYFRAME_STREAK: u32 = 3;
|
||||
|
||||
/// Longest the pump holds the last good frame waiting for a post-loss re-anchor keyframe before it
|
||||
/// gives up and resumes display. After a reference loss the hardware decoder does not error — it
|
||||
/// conceals the reference-missing deltas (on RADV, the DPB-and-output-COINCIDE path renders them as
|
||||
/// a gray plate with the new frame's motion painted over it) and returns Ok, so displaying them is
|
||||
/// the "gray frames mid-stream" artifact. We instead freeze on the last good picture until a fresh
|
||||
/// IDR re-anchors decode — the behaviour NVIDIA already shows (its DISTINCT output image + different
|
||||
/// concealment reads as a brief freeze, not gray). This cap only bounds the freeze when recovery
|
||||
/// genuinely stalls (host ignores the request, or an RFI recovery that never emits a keyframe), so a
|
||||
/// glitch can never become a permanent freeze. A recovery IDR round-trips well under this on any
|
||||
/// live link.
|
||||
const REANCHOR_FREEZE_MAX: Duration = Duration::from_millis(500);
|
||||
|
||||
/// How many host intra-refresh recovery marks ([`USER_FLAG_RECOVERY_POINT`]) must arrive since the
|
||||
/// latest frame gap before the pump lifts its freeze on an IDR-free stream. TWO, not one: with a
|
||||
/// continuous rolling wave the host marks phase-fixed wave boundaries, so the FIRST boundary after a
|
||||
/// loss is only partially healed — stripes swept BEFORE the loss still reference the lost frame — and
|
||||
/// lifting there would flash a partially-stale picture. The SECOND boundary guarantees a full wave
|
||||
/// swept entirely after the loss, so the picture is clean. This stays correct under repeated loss
|
||||
/// because every new gap resets the count. The cost is up to ~2 wave periods of holding the last good
|
||||
/// frame — the deliberate "hold longer, never show garbage" trade.
|
||||
///
|
||||
/// [`USER_FLAG_RECOVERY_POINT`]: punktfunk_core::packet::USER_FLAG_RECOVERY_POINT
|
||||
const REANCHOR_MARKS_TO_LIFT: u32 = 2;
|
||||
|
||||
/// Backstop patience while a host intra-refresh heal is visibly in progress. Each recovery mark
|
||||
/// pushes the freeze deadline out by this much, so a live mark stream (the host actively healing via
|
||||
/// its wave) keeps the client patiently holding the last good frame instead of tripping the IDR
|
||||
/// floor mid-heal. Must exceed the inter-mark interval (one wave period, ~0.5 s) with margin; if the
|
||||
/// marks STOP (heal stalled, or the host isn't running intra-refresh) the deadline lapses and the
|
||||
/// normal recovery-IDR floor fires, so a real stall still recovers.
|
||||
const RECOVERY_MARK_PATIENCE: Duration = Duration::from_millis(1500);
|
||||
|
||||
/// Frames skipped when `got` arrives while `expected` was the next index, or `None` if `got` is
|
||||
/// contiguous (`== expected`) or a straggler we have already passed. Frame indices are u32 counters
|
||||
/// that wrap, so the "ahead" test is a wrapping subtraction split at the half-space: a small
|
||||
/// positive delta is a forward gap (missing frames whose dependents will decode against absent
|
||||
/// references); a delta in the top half is an index behind us.
|
||||
fn index_gap(expected: u32, got: u32) -> Option<u32> {
|
||||
let ahead = got.wrapping_sub(expected);
|
||||
(ahead != 0 && ahead < u32::MAX / 2).then_some(ahead)
|
||||
}
|
||||
|
||||
/// Fold one decoded frame into the re-anchor state and decide whether it lifts the post-loss freeze.
|
||||
///
|
||||
/// `is_keyframe` — a real IDR (always a clean re-anchor). `has_anchor` — this AU carried
|
||||
/// [`USER_FLAG_RECOVERY_ANCHOR`](punktfunk_core::packet::USER_FLAG_RECOVERY_ANCHOR), the host's
|
||||
/// definitive single-frame re-anchor from an LTR-RFI recovery (a clean P-frame coded against a
|
||||
/// known-good reference), so it lifts on the FIRST occurrence exactly like an IDR — no two-mark wait.
|
||||
/// `has_mark` — this AU carried [`USER_FLAG_RECOVERY_POINT`](punktfunk_core::packet::USER_FLAG_RECOVERY_POINT),
|
||||
/// a host-signalled intra-refresh wave boundary (only *half* a re-anchor). `marks` — recovery marks
|
||||
/// seen since the latest gap.
|
||||
///
|
||||
/// Returns `(lift, new_marks)`: `lift` clears the freeze; `new_marks` is the running count (reset to 0
|
||||
/// on a lift). The two-mark rule ([`REANCHOR_MARKS_TO_LIFT`]) lives here so it is unit-tested
|
||||
/// independent of the pump's channel/decoder plumbing — the first wave boundary after a loss is only
|
||||
/// partially healed, so a single mark must NOT lift. An anchor (or IDR) is a *whole* re-anchor and
|
||||
/// lifts immediately.
|
||||
fn reanchor_after_frame(
|
||||
is_keyframe: bool,
|
||||
has_anchor: bool,
|
||||
has_mark: bool,
|
||||
marks: u32,
|
||||
) -> (bool, u32) {
|
||||
let marks = if has_mark {
|
||||
marks.saturating_add(1)
|
||||
} else {
|
||||
marks
|
||||
};
|
||||
if is_keyframe || has_anchor || marks >= REANCHOR_MARKS_TO_LIFT {
|
||||
(true, 0)
|
||||
} else {
|
||||
(false, marks)
|
||||
}
|
||||
}
|
||||
|
||||
/// Frames the pump keeps waiting for their 0xCF host timing (pts → capture→received µs).
|
||||
/// ~2 s at 120 Hz — a timing arrives within a frame or two of its AU, and against an old
|
||||
/// host (no 0xCF at all) this just caps the dead-weight ring.
|
||||
@@ -382,27 +303,17 @@ fn pump(
|
||||
// What actually decoded the last frame — a VAAPI failure demotes mid-session, so
|
||||
// this is read off each frame's image variant rather than fixed at startup.
|
||||
let mut dec_path: &'static str = "";
|
||||
// Loss recovery: watch the host→client unrecoverable-drop count and ask for an IDR when it climbs.
|
||||
let mut last_dropped = connector.frames_dropped();
|
||||
// The stats window keeps its own drop cursor — the OSD shows the per-window delta.
|
||||
let mut window_dropped = last_dropped;
|
||||
let mut window_dropped = connector.frames_dropped();
|
||||
let mut last_kf_req: Option<Instant> = None;
|
||||
// Consecutive received AUs that produced NO decoded frame (decode error, or the
|
||||
// decoder swallowed a reference-missing delta and returned nothing). Distinct from
|
||||
// `frames_dropped`, which counts reassembler drops: when the initial IDR is lost (or
|
||||
// we join mid-GOP) the reassembler delivers complete-but-undecodable deltas — it
|
||||
// never drops, so the drop-count trigger below stays silent and the stream freezes
|
||||
// on the last good frame. A short streak forces a fresh IDR to re-anchor.
|
||||
let mut no_output_streak = 0u32;
|
||||
// Freeze-until-reanchor: armed the moment we request a recovery keyframe (loss, decode error, or
|
||||
// a no-output streak), it withholds the decoder's concealed frames from the presenter — which
|
||||
// then redraws the last good picture — until a fresh keyframe re-anchors decode. See
|
||||
// [`REANCHOR_FREEZE_MAX`] for why this exists and its backstop deadline.
|
||||
let mut awaiting_reanchor = false;
|
||||
let mut reanchor_deadline: Option<Instant> = None;
|
||||
// Host intra-refresh recovery marks seen since the latest gap (see [`REANCHOR_MARKS_TO_LIFT`]).
|
||||
// Reset to 0 whenever the freeze is (re-)armed, so a fresh loss always waits out two fresh marks.
|
||||
let mut recovery_marks: u32 = 0;
|
||||
// Freeze-until-reanchor: the shared post-loss gate ([`punktfunk_core::reanchor::ReanchorGate`]).
|
||||
// Armed on any loss signal (frame-index gap, dropped-count climb, decoder wedge/demotion), it
|
||||
// withholds the decoder's concealed frames from the presenter — which then redraws the last good
|
||||
// picture — until a proven clean re-anchor (IDR / RFI anchor / second recovery mark) lifts it. It
|
||||
// also owns the no-output streak and the overdue-freeze backstop; the client keeps its own
|
||||
// `last_kf_req` request throttle and routes the gate's keyframe intents through it. Seeded with the
|
||||
// current drop count so the first `poll` doesn't read the baseline as a loss.
|
||||
let mut gate = ReanchorGate::new(connector.frames_dropped());
|
||||
// The frame_index we expect next (the host numbers frames consecutively). A jump means a frame
|
||||
// went missing — the earliest, most reliable signal that the decoder is about to conceal, ~120 ms
|
||||
// ahead of `frames_dropped` (the reassembler only declares a straggler lost once it ages out of
|
||||
@@ -447,9 +358,7 @@ fn pump(
|
||||
Some(exp) => {
|
||||
if let Some(gap) = index_gap(exp, frame.frame_index) {
|
||||
let now = Instant::now();
|
||||
awaiting_reanchor = true;
|
||||
recovery_marks = 0;
|
||||
reanchor_deadline = Some(now + REANCHOR_FREEZE_MAX);
|
||||
gate.arm(now);
|
||||
next_expected_index = Some(frame.frame_index.wrapping_add(1));
|
||||
// The gap carries the PRECISE lost range — [first missing, newest
|
||||
// received - 1] — so this is the one recovery signal that can drive true
|
||||
@@ -488,38 +397,14 @@ fn pump(
|
||||
}
|
||||
match decoder.decode(&frame.data) {
|
||||
Ok(Some(image)) => {
|
||||
// A decoded frame — the anchor holds.
|
||||
no_output_streak = 0;
|
||||
// Host-signalled intra-refresh recovery mark: on an IDR-free intra-refresh
|
||||
// stream this wave-boundary flag is the only clean point the client can honor
|
||||
// (the decoder never flags the re-anchor — the coded frame stays `P`). A live
|
||||
// mark stream also means the host is actively healing, so push the backstop out
|
||||
// rather than trip a mid-heal IDR (see `RECOVERY_MARK_PATIENCE`).
|
||||
let has_mark =
|
||||
frame.flags & punktfunk_core::packet::USER_FLAG_RECOVERY_POINT != 0;
|
||||
// The host's definitive single-frame re-anchor: an LTR-RFI recovery frame (a
|
||||
// clean P-frame off a known-good reference), the AMD twin of an IDR re-anchor
|
||||
// but without the spike. It lifts on the FIRST occurrence.
|
||||
let has_anchor =
|
||||
frame.flags & punktfunk_core::packet::USER_FLAG_RECOVERY_ANCHOR != 0;
|
||||
if has_mark && awaiting_reanchor {
|
||||
reanchor_deadline = Some(Instant::now() + RECOVERY_MARK_PATIENCE);
|
||||
}
|
||||
// A fresh clean re-anchor lifts the freeze and shows this frame: a real intra
|
||||
// keyframe (IDR, always clean), an LTR-RFI recovery anchor (also whole), OR the
|
||||
// second recovery mark since the gap (the first wave boundary is only
|
||||
// half-healed — see `reanchor_after_frame`).
|
||||
let (lift, marks) = reanchor_after_frame(
|
||||
image.is_keyframe(),
|
||||
has_anchor,
|
||||
has_mark,
|
||||
recovery_marks,
|
||||
);
|
||||
recovery_marks = marks;
|
||||
if lift {
|
||||
awaiting_reanchor = false;
|
||||
reanchor_deadline = None;
|
||||
}
|
||||
// Fold this decoded frame through the shared freeze gate: it reads the AU's
|
||||
// re-anchor wire flags (FLAG_SOF IDR marker / RECOVERY_ANCHOR / RECOVERY_POINT),
|
||||
// takes `image.is_keyframe()` as the ffmpeg keyframe belt, applies the two-mark
|
||||
// rule + the mark-patience backstop, clears the no-output streak, and returns
|
||||
// whether to present this frame or withhold it as a post-loss concealment.
|
||||
let present = gate
|
||||
.on_decoded(frame.flags, image.is_keyframe(), Instant::now())
|
||||
== GateVerdict::Present;
|
||||
total_frames += 1;
|
||||
dec_path = match &image {
|
||||
DecodedImage::Cpu(_) => "software",
|
||||
@@ -574,19 +459,19 @@ fn pump(
|
||||
DecodedImage::VkFrame(v) => Some((v.timeline_sem, v.decode_done_value)),
|
||||
_ => None,
|
||||
};
|
||||
if awaiting_reanchor {
|
||||
// Post-loss concealment: withhold this frame (it references a lost/gray
|
||||
// reference) so the presenter keeps redrawing the last good picture
|
||||
// rather than flashing the decoder's gray plate. Dropped here — the
|
||||
// hw-decode stat below still samples via `hw_fence` (raw handle + value,
|
||||
// valid past the guard). Cleared by the next keyframe or the backstop.
|
||||
tracing::trace!("holding last frame — awaiting post-loss re-anchor");
|
||||
} else {
|
||||
if present {
|
||||
let _ = frame_tx.force_send(DecodedFrame {
|
||||
pts_ns: frame.pts_ns,
|
||||
decoded_ns,
|
||||
image,
|
||||
});
|
||||
} else {
|
||||
// Post-loss concealment: withhold this frame (it references a lost/gray
|
||||
// reference) so the presenter keeps redrawing the last good picture rather
|
||||
// than flashing the decoder's gray plate. Dropped here — the hw-decode stat
|
||||
// below still samples via `hw_fence` (raw handle + value, valid past the
|
||||
// guard). The gate lifts the freeze on the next clean re-anchor / backstop.
|
||||
tracing::trace!("holding last frame — awaiting post-loss re-anchor");
|
||||
}
|
||||
// `decode` stage: received→decode COMPLETE, single clock.
|
||||
match hw_fence {
|
||||
@@ -602,36 +487,35 @@ fn pump(
|
||||
}
|
||||
}
|
||||
}
|
||||
Ok(None) => no_output_streak += 1,
|
||||
// Survivable (loss until the next IDR/RFI recovery) — keep feeding.
|
||||
Err(e) => {
|
||||
no_output_streak += 1;
|
||||
tracing::debug!(error = %e, "decode error (recovering)");
|
||||
}
|
||||
}
|
||||
// The decoder has produced nothing for a short run — under zero-reorder
|
||||
// LOW_DELAY (one-in/one-out) that means it's wedged on missing references
|
||||
// with no reassembler drop to trigger recovery below. Ask for a fresh IDR
|
||||
// (throttled), then re-arm the streak so we wait out the request→IDR round
|
||||
// trip before asking again instead of flooding.
|
||||
if no_output_streak >= NO_OUTPUT_KEYFRAME_STREAK {
|
||||
// The decoder produced nothing — under zero-reorder LOW_DELAY (one-in/one-out) that
|
||||
// means it's wedged on missing references with no reassembler drop to trigger
|
||||
// recovery. The gate counts the streak and, once it trips, arms the freeze and tells
|
||||
// us to (throttled) request a fresh IDR to re-anchor. Both the empty-output and the
|
||||
// survivable-decode-error arms feed it; a decoded frame resets the streak in
|
||||
// `on_decoded`.
|
||||
Ok(None) => {
|
||||
let now = Instant::now();
|
||||
// Wedged on missing references: hold the last good frame until re-anchor
|
||||
// (armed even when the IDR request itself is throttled — the stream is broken
|
||||
// regardless of whether we ask again this iteration).
|
||||
awaiting_reanchor = true;
|
||||
recovery_marks = 0;
|
||||
reanchor_deadline = Some(now + REANCHOR_FREEZE_MAX);
|
||||
if last_kf_req
|
||||
if gate.on_no_output(now)
|
||||
&& last_kf_req
|
||||
.is_none_or(|t| now.duration_since(t) >= Duration::from_millis(100))
|
||||
{
|
||||
last_kf_req = Some(now);
|
||||
let _ = connector.request_keyframe();
|
||||
tracing::debug!(
|
||||
streak = no_output_streak,
|
||||
"requested keyframe (decoder produced no output)"
|
||||
);
|
||||
no_output_streak = 0;
|
||||
tracing::debug!("requested keyframe (decoder produced no output)");
|
||||
}
|
||||
}
|
||||
// Survivable (loss until the next IDR/RFI recovery) — keep feeding.
|
||||
Err(e) => {
|
||||
tracing::debug!(error = %e, "decode error (recovering)");
|
||||
let now = Instant::now();
|
||||
if gate.on_no_output(now)
|
||||
&& last_kf_req
|
||||
.is_none_or(|t| now.duration_since(t) >= Duration::from_millis(100))
|
||||
{
|
||||
last_kf_req = Some(now);
|
||||
let _ = connector.request_keyframe();
|
||||
tracing::debug!("requested keyframe (decode error recovery)");
|
||||
}
|
||||
}
|
||||
}
|
||||
// The presenter's verdict: hardware frames can't be displayed (GL converter
|
||||
@@ -649,9 +533,7 @@ fn pump(
|
||||
// through the same throttle as loss recovery below.
|
||||
if decoder.take_keyframe_request() {
|
||||
let now = Instant::now();
|
||||
awaiting_reanchor = true;
|
||||
recovery_marks = 0;
|
||||
reanchor_deadline = Some(now + REANCHOR_FREEZE_MAX);
|
||||
gate.arm(now);
|
||||
if last_kf_req
|
||||
.is_none_or(|t| now.duration_since(t) >= Duration::from_millis(100))
|
||||
{
|
||||
@@ -679,41 +561,23 @@ fn pump(
|
||||
}
|
||||
}
|
||||
|
||||
// Loss recovery: under infinite GOP the only recovery keyframe is one we request. The
|
||||
// reassembler drops unrecoverable AUs (frames_dropped); the decoder then conceals the
|
||||
// reference-missing delta frames that follow and returns Ok, so keying off a decode error
|
||||
// rarely fires. Request an IDR when the drop count climbs, throttled — the decode stays
|
||||
// wedged for several frames until the IDR lands, so requesting every frame would flood.
|
||||
// Loss recovery + overdue backstop, folded through the shared gate. A climb in the
|
||||
// reassembler's unrecoverable-drop count (`frames_dropped`) means the AUs after the lost one
|
||||
// reference a picture we never decoded — the decoder conceals them (gray on RADV) and returns
|
||||
// Ok, so a decode-error trigger rarely fires; the gate arms the freeze on the climb instead. An
|
||||
// overdue freeze (held a full REANCHOR_FREEZE_MAX with no clean re-anchor — a lost recovery IDR,
|
||||
// or a benign reorder that produced no `frames_dropped`) re-asks while it keeps holding: NEVER
|
||||
// resume to gray — a genuinely dead stream is the QUIC idle-timeout watchdog's job. Both route
|
||||
// the gate's keyframe intent through the shared 100 ms throttle; under infinite GOP the only
|
||||
// recovery keyframe is one we request.
|
||||
let dropped = connector.frames_dropped();
|
||||
if dropped > last_dropped {
|
||||
last_dropped = dropped;
|
||||
let now = Instant::now();
|
||||
// A dropped AU means the frames after it reference a picture we never decoded — the
|
||||
// decoder will conceal them (gray on RADV). Freeze on the last good frame until a fresh
|
||||
// IDR re-anchors, so the concealment never reaches the screen.
|
||||
awaiting_reanchor = true;
|
||||
recovery_marks = 0;
|
||||
reanchor_deadline = Some(now + REANCHOR_FREEZE_MAX);
|
||||
if last_kf_req.is_none_or(|t| now.duration_since(t) >= Duration::from_millis(100)) {
|
||||
if gate.poll(dropped, now)
|
||||
&& last_kf_req.is_none_or(|t| now.duration_since(t) >= Duration::from_millis(100))
|
||||
{
|
||||
last_kf_req = Some(now);
|
||||
let _ = connector.request_keyframe();
|
||||
tracing::debug!(dropped, "requested keyframe (loss recovery)");
|
||||
}
|
||||
}
|
||||
// Re-anchor overdue: the freeze has held the whole window with no keyframe — a lost recovery
|
||||
// IDR, or a benign reorder that produced no `frames_dropped` and so requested none. Do NOT
|
||||
// resume to gray (the one thing worse than a freeze): keep holding the last good frame and
|
||||
// (re-)request a keyframe, throttled + host-coalesced, so a CLEAN re-anchor is what un-freezes
|
||||
// us. A genuinely dead stream — host gone, link collapsed — is caught by the QUIC idle-timeout
|
||||
// watchdog (returns to the menu), never by painting the decoder's concealment.
|
||||
if awaiting_reanchor && reanchor_deadline.is_some_and(|d| Instant::now() >= d) {
|
||||
let now = Instant::now();
|
||||
reanchor_deadline = Some(now + REANCHOR_FREEZE_MAX);
|
||||
if last_kf_req.is_none_or(|t| now.duration_since(t) >= Duration::from_millis(100)) {
|
||||
last_kf_req = Some(now);
|
||||
let _ = connector.request_keyframe();
|
||||
tracing::debug!("re-anchor overdue — still holding, re-requesting keyframe");
|
||||
}
|
||||
tracing::debug!(dropped, "requested keyframe (loss recovery / overdue re-anchor)");
|
||||
}
|
||||
|
||||
if window_start.elapsed() >= Duration::from_secs(1) {
|
||||
@@ -836,111 +700,3 @@ fn spawn_audio(
|
||||
.map_err(|e| tracing::warn!(error = %e, "audio thread failed to start — audio disabled"))
|
||||
.ok()
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::{index_gap, reanchor_after_frame, REANCHOR_MARKS_TO_LIFT};
|
||||
|
||||
// Simulate the pump's re-anchor state across a sequence of decoded frames: each `(is_keyframe,
|
||||
// has_mark)` pair is folded through `reanchor_after_frame`, returning the frame index (0-based)
|
||||
// at which the freeze first lifts, or `None` if it never does. `gap_before` reset points model a
|
||||
// fresh loss re-arming the freeze (the pump zeroes the count at every gap/arm site).
|
||||
fn lift_at(frames: &[(bool, bool)]) -> Option<usize> {
|
||||
let mut marks = 0u32;
|
||||
for (i, &(is_kf, has_mark)) in frames.iter().enumerate() {
|
||||
// The intra-refresh-mark model never carries an LTR-RFI anchor (that path is exercised
|
||||
// by `an_rfi_anchor_lifts_immediately`), so `has_anchor` is always false here.
|
||||
let (lift, m) = reanchor_after_frame(is_kf, false, has_mark, marks);
|
||||
marks = m;
|
||||
if lift {
|
||||
return Some(i);
|
||||
}
|
||||
}
|
||||
None
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_single_recovery_mark_does_not_lift() {
|
||||
// The first wave boundary after a loss is only half-healed — one mark must hold the freeze.
|
||||
assert_eq!(REANCHOR_MARKS_TO_LIFT, 2);
|
||||
assert_eq!(lift_at(&[(false, true)]), None);
|
||||
assert_eq!(
|
||||
lift_at(&[(false, false), (false, true), (false, false)]),
|
||||
None
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn the_second_recovery_mark_lifts() {
|
||||
// Two marks = a full wave swept after the loss → clean re-anchor.
|
||||
assert_eq!(lift_at(&[(false, true), (false, true)]), Some(1));
|
||||
assert_eq!(
|
||||
lift_at(&[(false, false), (false, true), (false, false), (false, true)]),
|
||||
Some(3)
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_real_keyframe_lifts_immediately() {
|
||||
// An IDR is always a clean anchor — no marks needed.
|
||||
assert_eq!(lift_at(&[(true, false)]), Some(0));
|
||||
assert_eq!(lift_at(&[(false, true), (true, false)]), Some(1));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_fresh_gap_resets_the_mark_count() {
|
||||
// The pump zeroes `recovery_marks` at each arm site, so one mark before a new gap plus one
|
||||
// after must NOT lift — the model resets the running count to imitate that.
|
||||
let mut marks = 0u32;
|
||||
let (_, m) = reanchor_after_frame(false, false, true, marks); // mark #1 (pre-gap)
|
||||
marks = m;
|
||||
assert_eq!(marks, 1);
|
||||
marks = 0; // a new gap re-arms the freeze → count reset
|
||||
let (lift, m) = reanchor_after_frame(false, false, true, marks); // first mark of the new wave
|
||||
assert!(!lift, "a single post-gap mark must not lift");
|
||||
assert_eq!(m, 1);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn an_rfi_anchor_lifts_immediately() {
|
||||
// An LTR-RFI recovery anchor is a WHOLE re-anchor (a clean P-frame off a known-good
|
||||
// reference), so — like an IDR — it lifts on the FIRST occurrence, no two-mark wait.
|
||||
let (lift, marks) = reanchor_after_frame(false, true, false, 0);
|
||||
assert!(lift, "an RFI anchor must lift the freeze immediately");
|
||||
assert_eq!(marks, 0, "a lift resets the running mark count");
|
||||
// Even with zero prior marks and no keyframe, the anchor alone is sufficient.
|
||||
let (lift, _) = reanchor_after_frame(false, true, true, 1);
|
||||
assert!(lift, "an anchor lifts regardless of the pending mark count");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn contiguous_indices_are_not_a_gap() {
|
||||
assert_eq!(index_gap(5, 5), None);
|
||||
assert_eq!(index_gap(0, 0), None);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_forward_jump_reports_the_skip_count() {
|
||||
assert_eq!(index_gap(5, 6), Some(1)); // one frame missing
|
||||
assert_eq!(index_gap(5, 9), Some(4));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_straggler_behind_us_is_not_a_gap() {
|
||||
// The reassembler emitted a newer frame first; the late one must not re-arm.
|
||||
assert_eq!(index_gap(9, 5), None);
|
||||
assert_eq!(index_gap(1, 0), None);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn the_index_counter_wraps_cleanly() {
|
||||
// last frame = u32::MAX, so the next expected wraps to 0.
|
||||
// Contiguous across the wrap.
|
||||
assert_eq!(index_gap(0, 0), None);
|
||||
// waiting on u32::MAX, frame 0 arrived → MAX was skipped.
|
||||
assert_eq!(index_gap(u32::MAX, 0), Some(1));
|
||||
assert_eq!(index_gap(u32::MAX, 2), Some(3));
|
||||
// an old frame arriving just after the wrap is still a straggler.
|
||||
assert_eq!(index_gap(0, u32::MAX), None);
|
||||
}
|
||||
}
|
||||
|
||||
@@ -13,6 +13,7 @@
|
||||
use crate::config::{Config, FecConfig, FecScheme, ProtocolPhase, Role};
|
||||
use crate::error::PunktfunkStatus;
|
||||
use crate::input::InputEvent;
|
||||
use crate::reanchor::{GateVerdict, ReanchorGate};
|
||||
use crate::session::Session;
|
||||
use crate::stats::Stats;
|
||||
use crate::transport::{loopback_pair, Transport, UdpTransport};
|
||||
@@ -2620,3 +2621,142 @@ pub unsafe extern "C" fn punktfunk_connection_close(c: *mut PunktfunkConnection)
|
||||
drop(unsafe { Box::from_raw(c) });
|
||||
}
|
||||
}
|
||||
|
||||
// ---- Post-loss re-anchor freeze gate ----
|
||||
//
|
||||
// The shared [`ReanchorGate`](crate::reanchor::ReanchorGate) exposed for the Swift client (Rust
|
||||
// embedders — Android/Windows/Linux — use the struct directly). After an unrecoverable reference
|
||||
// loss the decoder silently conceals the missing-reference deltas (gray/garbage picture, no error);
|
||||
// the client freezes on the last good frame and lifts only on a proven clean re-anchor. The gate
|
||||
// takes time internally (`Instant::now`) so no timestamps cross the boundary. Drive it per session:
|
||||
// `arm` on a loss (frame-index gap from `punktfunk_connection_note_frame_index`, a decoder
|
||||
// wedge/demotion), `on_decoded` per decoded frame to gate presentation, `on_no_output` per AU that
|
||||
// produced nothing, and `poll` each iteration for the dropped-count climb + overdue backstop. Route
|
||||
// the returned keyframe intents through the client's existing request throttle.
|
||||
|
||||
/// Create a re-anchor gate seeded with the session's current `frames_dropped` (so the first
|
||||
/// [`punktfunk_reanchor_gate_poll`] doesn't read the baseline as a loss). Free with
|
||||
/// [`punktfunk_reanchor_gate_free`]. Never returns NULL.
|
||||
#[no_mangle]
|
||||
pub extern "C" fn punktfunk_reanchor_gate_new(frames_dropped: u64) -> *mut ReanchorGate {
|
||||
Box::into_raw(Box::new(ReanchorGate::new(frames_dropped)))
|
||||
}
|
||||
|
||||
/// Free a gate created by [`punktfunk_reanchor_gate_new`]. NULL is a no-op.
|
||||
///
|
||||
/// # Safety
|
||||
/// `g` was returned by [`punktfunk_reanchor_gate_new`] and is not used after this call.
|
||||
#[no_mangle]
|
||||
pub unsafe extern "C" fn punktfunk_reanchor_gate_free(g: *mut ReanchorGate) {
|
||||
if !g.is_null() {
|
||||
drop(unsafe { Box::from_raw(g) });
|
||||
}
|
||||
}
|
||||
|
||||
/// Arm the freeze: a loss was detected (a frame-index gap, or a decoder wedge/demotion). Zeroes the
|
||||
/// recovery-mark count and (re-)sets the backstop deadline. NULL is a no-op.
|
||||
///
|
||||
/// # Safety
|
||||
/// `g` is a valid gate handle.
|
||||
#[no_mangle]
|
||||
pub unsafe extern "C" fn punktfunk_reanchor_gate_arm(g: *mut ReanchorGate) {
|
||||
if let Some(g) = unsafe { g.as_mut() } {
|
||||
g.arm(std::time::Instant::now());
|
||||
}
|
||||
}
|
||||
|
||||
/// Fold one decoded frame and write to `out_present` whether to display it (`true`) or withhold it as
|
||||
/// a post-loss concealment (`false`). `flags` is the AU's `user_flags` word ([`PunktfunkFrame::flags`]):
|
||||
/// the gate reads `FLAG_SOF` (the host's IDR marker), `USER_FLAG_RECOVERY_ANCHOR` and
|
||||
/// `USER_FLAG_RECOVERY_POINT`. Pass `decoder_keyframe = false` where the platform decoder doesn't flag
|
||||
/// IDRs (VideoToolbox/MediaCodec) — the wire `FLAG_SOF` covers it.
|
||||
///
|
||||
/// # Safety
|
||||
/// `g` is a valid gate handle; `out_present` is writable or NULL.
|
||||
#[no_mangle]
|
||||
pub unsafe extern "C" fn punktfunk_reanchor_gate_on_decoded(
|
||||
g: *mut ReanchorGate,
|
||||
flags: u32,
|
||||
decoder_keyframe: bool,
|
||||
out_present: *mut bool,
|
||||
) -> PunktfunkStatus {
|
||||
guard(|| {
|
||||
let g = match unsafe { g.as_mut() } {
|
||||
Some(g) => g,
|
||||
None => return PunktfunkStatus::NullPointer,
|
||||
};
|
||||
let present = g.on_decoded(flags, decoder_keyframe, std::time::Instant::now()) == GateVerdict::Present;
|
||||
if !out_present.is_null() {
|
||||
unsafe { *out_present = present };
|
||||
}
|
||||
PunktfunkStatus::Ok
|
||||
})
|
||||
}
|
||||
|
||||
/// A received AU produced no decoded frame. Writes to `out_request_kf` whether the no-output streak has
|
||||
/// tripped and the client should (throttled) request a keyframe — the gate arms the freeze at the same
|
||||
/// time.
|
||||
///
|
||||
/// # Safety
|
||||
/// `g` is a valid gate handle; `out_request_kf` is writable or NULL.
|
||||
#[no_mangle]
|
||||
pub unsafe extern "C" fn punktfunk_reanchor_gate_on_no_output(
|
||||
g: *mut ReanchorGate,
|
||||
out_request_kf: *mut bool,
|
||||
) -> PunktfunkStatus {
|
||||
guard(|| {
|
||||
let g = match unsafe { g.as_mut() } {
|
||||
Some(g) => g,
|
||||
None => return PunktfunkStatus::NullPointer,
|
||||
};
|
||||
let request = g.on_no_output(std::time::Instant::now());
|
||||
if !out_request_kf.is_null() {
|
||||
unsafe { *out_request_kf = request };
|
||||
}
|
||||
PunktfunkStatus::Ok
|
||||
})
|
||||
}
|
||||
|
||||
/// Periodic fold of the session's `frames_dropped` counter plus the overdue backstop. Writes to
|
||||
/// `out_request_kf` whether the client should (throttled) request a keyframe (a drop-count climb armed
|
||||
/// a fresh freeze, or the freeze is overdue and re-asks while it keeps holding).
|
||||
///
|
||||
/// # Safety
|
||||
/// `g` is a valid gate handle; `out_request_kf` is writable or NULL.
|
||||
#[no_mangle]
|
||||
pub unsafe extern "C" fn punktfunk_reanchor_gate_poll(
|
||||
g: *mut ReanchorGate,
|
||||
frames_dropped: u64,
|
||||
out_request_kf: *mut bool,
|
||||
) -> PunktfunkStatus {
|
||||
guard(|| {
|
||||
let g = match unsafe { g.as_mut() } {
|
||||
Some(g) => g,
|
||||
None => return PunktfunkStatus::NullPointer,
|
||||
};
|
||||
let request = g.poll(frames_dropped, std::time::Instant::now());
|
||||
if !out_request_kf.is_null() {
|
||||
unsafe { *out_request_kf = request };
|
||||
}
|
||||
PunktfunkStatus::Ok
|
||||
})
|
||||
}
|
||||
|
||||
/// Whether the gate is currently withholding concealed frames (frozen on the last good picture).
|
||||
/// Writes `false` on a NULL gate.
|
||||
///
|
||||
/// # Safety
|
||||
/// `g` is a valid gate handle; `out_holding` is writable or NULL.
|
||||
#[no_mangle]
|
||||
pub unsafe extern "C" fn punktfunk_reanchor_gate_is_holding(
|
||||
g: *const ReanchorGate,
|
||||
out_holding: *mut bool,
|
||||
) -> PunktfunkStatus {
|
||||
guard(|| {
|
||||
let holding = unsafe { g.as_ref() }.is_some_and(ReanchorGate::is_holding);
|
||||
if !out_holding.is_null() {
|
||||
unsafe { *out_holding = holding };
|
||||
}
|
||||
PunktfunkStatus::Ok
|
||||
})
|
||||
}
|
||||
|
||||
@@ -38,6 +38,7 @@ pub mod input;
|
||||
pub mod packet;
|
||||
#[cfg(feature = "quic")]
|
||||
pub mod quic;
|
||||
pub mod reanchor;
|
||||
pub mod session;
|
||||
pub mod stats;
|
||||
pub mod transport;
|
||||
@@ -61,7 +62,10 @@ pub use stats::Stats;
|
||||
/// TTL of a v2 envelope; `punktfunk_connection_next_rumble` is unchanged and drops it). Additive —
|
||||
/// the wire is backward-compatible (the envelope is a length-tolerant tail on 0xCA), so
|
||||
/// [`WIRE_VERSION`] is unchanged.
|
||||
pub const ABI_VERSION: u32 = 5;
|
||||
/// v6: added the `punktfunk_reanchor_gate_*` surface (post-loss freeze-until-reanchor gate for the
|
||||
/// Swift client; Rust embedders use [`reanchor::ReanchorGate`] directly). Additive, client-local —
|
||||
/// no wire change, so [`WIRE_VERSION`] is unchanged.
|
||||
pub const ABI_VERSION: u32 = 6;
|
||||
|
||||
/// The punktfunk/1 **wire** version — what `Hello`/`Welcome` carry and hosts equality-check.
|
||||
/// Deliberately its own constant: [`ABI_VERSION`] tracks the embeddable **C surface**
|
||||
|
||||
@@ -0,0 +1,456 @@
|
||||
//! Post-loss display freeze — the shared "freeze-until-reanchor" gate.
|
||||
//!
|
||||
//! After an unrecoverable reference loss the hardware decoder does **not** error: it *conceals* the
|
||||
//! reference-missing delta frames (on RADV, the DPB-and-output-COINCIDE path paints a gray plate with
|
||||
//! the new frame's motion on top) and returns Ok. Displaying that is the "gray frames mid-stream"
|
||||
//! artifact. Instead every client freezes on the last good picture — withholds the concealed frames
|
||||
//! from its presenter, which keeps redrawing the held frame — and lifts the freeze ONLY on a proven
|
||||
//! clean re-anchor: a real IDR, an LTR-RFI recovery anchor ([`USER_FLAG_RECOVERY_ANCHOR`]), or the
|
||||
//! second intra-refresh recovery mark ([`USER_FLAG_RECOVERY_POINT`]) since the loss.
|
||||
//!
|
||||
//! This module owns that decision so every embedder shares ONE implementation instead of re-deriving
|
||||
//! it (the Linux/Deck pump in `pf-client-core`, the Windows in-process pump, the Android decode loops,
|
||||
//! and — over the C ABI — the Apple client). The state machine is time-driven but takes `now` as a
|
||||
//! parameter so it is unit-testable without a clock; the C ABI wrappers supply `Instant::now()`.
|
||||
//!
|
||||
//! [`USER_FLAG_RECOVERY_POINT`]: crate::packet::USER_FLAG_RECOVERY_POINT
|
||||
//! [`USER_FLAG_RECOVERY_ANCHOR`]: crate::packet::USER_FLAG_RECOVERY_ANCHOR
|
||||
|
||||
use crate::packet::{FLAG_SOF, USER_FLAG_RECOVERY_ANCHOR, USER_FLAG_RECOVERY_POINT};
|
||||
use std::time::{Duration, Instant};
|
||||
|
||||
/// Consecutive no-output AUs that force a keyframe request. ~50 ms at 60 Hz — long enough not to fire
|
||||
/// on a one-frame decoder hiccup, short enough that a lost initial IDR (or a mid-GOP join) unfreezes
|
||||
/// almost immediately instead of never.
|
||||
pub const NO_OUTPUT_KEYFRAME_STREAK: u32 = 3;
|
||||
|
||||
/// Longest the gate holds the last good frame waiting for a post-loss re-anchor keyframe before it
|
||||
/// re-asks. After a reference loss the hardware decoder does not error — it conceals the
|
||||
/// reference-missing deltas (on RADV, the DPB-and-output-COINCIDE path renders them as a gray plate
|
||||
/// with the new frame's motion painted over it) and returns Ok, so displaying them is the "gray frames
|
||||
/// mid-stream" artifact. We instead freeze on the last good picture until a fresh IDR re-anchors decode
|
||||
/// — the behaviour NVIDIA already shows (its DISTINCT output image + different concealment reads as a
|
||||
/// brief freeze, not gray). This cap only bounds the freeze when recovery genuinely stalls (host
|
||||
/// ignores the request, or an RFI recovery that never emits a keyframe): the freeze is NEVER lifted to
|
||||
/// the concealed picture — the deadline re-asks for a keyframe and keeps holding, so a glitch can never
|
||||
/// become a permanent freeze while a clean re-anchor is what un-freezes. A recovery IDR round-trips well
|
||||
/// under this on any live link.
|
||||
pub const REANCHOR_FREEZE_MAX: Duration = Duration::from_millis(500);
|
||||
|
||||
/// How many host intra-refresh recovery marks ([`USER_FLAG_RECOVERY_POINT`]) must arrive since the
|
||||
/// latest loss before the gate lifts its freeze on an IDR-free stream. TWO, not one: with a continuous
|
||||
/// rolling wave the host marks phase-fixed wave boundaries, so the FIRST boundary after a loss is only
|
||||
/// partially healed — stripes swept BEFORE the loss still reference the lost frame — and lifting there
|
||||
/// would flash a partially-stale picture. The SECOND boundary guarantees a full wave swept entirely
|
||||
/// after the loss, so the picture is clean. This stays correct under repeated loss because every fresh
|
||||
/// arm resets the count. The cost is up to ~2 wave periods of holding the last good frame — the
|
||||
/// deliberate "hold longer, never show garbage" trade.
|
||||
///
|
||||
/// [`USER_FLAG_RECOVERY_POINT`]: crate::packet::USER_FLAG_RECOVERY_POINT
|
||||
pub const REANCHOR_MARKS_TO_LIFT: u32 = 2;
|
||||
|
||||
/// Backstop patience while a host intra-refresh heal is visibly in progress. Each recovery mark pushes
|
||||
/// the freeze deadline out by this much, so a live mark stream (the host actively healing via its wave)
|
||||
/// keeps the gate patiently holding the last good frame instead of tripping the IDR floor mid-heal.
|
||||
/// Must exceed the inter-mark interval (one wave period, ~0.5 s) with margin; if the marks STOP (heal
|
||||
/// stalled, or the host isn't running intra-refresh) the deadline lapses and the normal recovery-IDR
|
||||
/// floor fires, so a real stall still recovers.
|
||||
pub const RECOVERY_MARK_PATIENCE: Duration = Duration::from_millis(1500);
|
||||
|
||||
/// Frames skipped when `got` arrives while `expected` was the next index, or `None` if `got` is
|
||||
/// contiguous (`== expected`) or a straggler we have already passed. Frame indices are u32 counters
|
||||
/// that wrap, so the "ahead" test is a wrapping subtraction split at the half-space: a small positive
|
||||
/// delta is a forward gap (missing frames whose dependents will decode against absent references); a
|
||||
/// delta in the top half is an index behind us.
|
||||
pub fn index_gap(expected: u32, got: u32) -> Option<u32> {
|
||||
let ahead = got.wrapping_sub(expected);
|
||||
(ahead != 0 && ahead < u32::MAX / 2).then_some(ahead)
|
||||
}
|
||||
|
||||
/// Fold one decoded frame into the re-anchor state and decide whether it lifts the post-loss freeze.
|
||||
///
|
||||
/// `is_keyframe` — a real IDR (always a clean re-anchor). `has_anchor` — this AU carried
|
||||
/// [`USER_FLAG_RECOVERY_ANCHOR`](crate::packet::USER_FLAG_RECOVERY_ANCHOR), the host's definitive
|
||||
/// single-frame re-anchor from an LTR-RFI recovery (a clean P-frame coded against a known-good
|
||||
/// reference), so it lifts on the FIRST occurrence exactly like an IDR — no two-mark wait. `has_mark` —
|
||||
/// this AU carried [`USER_FLAG_RECOVERY_POINT`](crate::packet::USER_FLAG_RECOVERY_POINT), a
|
||||
/// host-signalled intra-refresh wave boundary (only *half* a re-anchor). `marks` — recovery marks seen
|
||||
/// since the latest loss.
|
||||
///
|
||||
/// Returns `(lift, new_marks)`: `lift` clears the freeze; `new_marks` is the running count (reset to 0
|
||||
/// on a lift). The two-mark rule ([`REANCHOR_MARKS_TO_LIFT`]) lives here so it is unit-tested
|
||||
/// independent of the pump's channel/decoder plumbing — the first wave boundary after a loss is only
|
||||
/// partially healed, so a single mark must NOT lift. An anchor (or IDR) is a *whole* re-anchor and
|
||||
/// lifts immediately.
|
||||
fn reanchor_after_frame(is_keyframe: bool, has_anchor: bool, has_mark: bool, marks: u32) -> (bool, u32) {
|
||||
let marks = if has_mark {
|
||||
marks.saturating_add(1)
|
||||
} else {
|
||||
marks
|
||||
};
|
||||
if is_keyframe || has_anchor || marks >= REANCHOR_MARKS_TO_LIFT {
|
||||
(true, 0)
|
||||
} else {
|
||||
(false, marks)
|
||||
}
|
||||
}
|
||||
|
||||
/// Whether a decoded frame should be shown or withheld while the gate is (or isn't) frozen.
|
||||
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
|
||||
pub enum GateVerdict {
|
||||
/// Present this frame — the gate is not frozen, or this frame is the clean re-anchor that lifts it.
|
||||
Present,
|
||||
/// Withhold this frame — it is a post-loss concealment; the presenter keeps the last good picture.
|
||||
Hold,
|
||||
}
|
||||
|
||||
/// The shared post-loss freeze state machine. A client feeds it three kinds of event — an *arm* (a
|
||||
/// loss was detected: a frame-index gap, a dropped-count climb, or a decoder wedge/demotion), each
|
||||
/// *decoded frame* ([`on_decoded`](Self::on_decoded), which decides present-vs-hold and interprets the
|
||||
/// re-anchor wire flags), and each *no-output* AU ([`on_no_output`](Self::on_no_output)) — plus a
|
||||
/// periodic [`poll`](Self::poll) that folds the dropped counter and fires the overdue backstop.
|
||||
///
|
||||
/// The gate emits *intents* only: [`on_no_output`](Self::on_no_output) and [`poll`](Self::poll) return
|
||||
/// `true` when the client should ask the host for a keyframe. The client routes that through its own
|
||||
/// ~100 ms request throttle (and the precise RFI-vs-keyframe range decision stays in the loss-range
|
||||
/// tracker behind [`crate::client::NativeClient::note_frame_index`]) — the gate never touches the wire.
|
||||
#[derive(Debug, Clone)]
|
||||
pub struct ReanchorGate {
|
||||
/// Frozen on the last good frame, withholding the decoder's concealed output until a clean
|
||||
/// re-anchor. Armed by any loss signal; cleared only by [`on_decoded`](Self::on_decoded) lifting.
|
||||
awaiting: bool,
|
||||
/// Host intra-refresh recovery marks seen since the latest arm (see [`REANCHOR_MARKS_TO_LIFT`]).
|
||||
/// Reset to 0 whenever the freeze is (re-)armed, so a fresh loss always waits out two fresh marks.
|
||||
marks: u32,
|
||||
/// When the freeze becomes overdue and [`poll`](Self::poll) re-asks for a keyframe (holding, never
|
||||
/// resuming to the concealed picture). `None` when not frozen.
|
||||
deadline: Option<Instant>,
|
||||
/// Consecutive received AUs that produced no decoded frame — a decoder wedged on missing references
|
||||
/// with no reassembler drop to trigger recovery. A short streak forces a fresh IDR.
|
||||
no_output_streak: u32,
|
||||
/// The last `frames_dropped` value [`poll`](Self::poll) observed; a climb means the reassembler
|
||||
/// declared an AU unrecoverable and the following deltas will conceal, so arm.
|
||||
last_dropped: u64,
|
||||
}
|
||||
|
||||
impl ReanchorGate {
|
||||
/// Seed the gate with the session's current `frames_dropped` so the first [`poll`](Self::poll)
|
||||
/// doesn't read the baseline as a loss.
|
||||
pub fn new(frames_dropped: u64) -> Self {
|
||||
ReanchorGate {
|
||||
awaiting: false,
|
||||
marks: 0,
|
||||
deadline: None,
|
||||
no_output_streak: 0,
|
||||
last_dropped: frames_dropped,
|
||||
}
|
||||
}
|
||||
|
||||
/// Arm the freeze: a loss was detected (a frame-index gap, a dropped-count climb, or a decoder
|
||||
/// wedge/demotion). Zeroes the mark count so a fresh loss waits out two fresh recovery marks, and
|
||||
/// (re-)sets the backstop deadline. Idempotent while already frozen (re-arming just re-zeroes the
|
||||
/// marks and pushes the deadline — the correct behaviour when a second loss lands mid-freeze).
|
||||
pub fn arm(&mut self, now: Instant) {
|
||||
self.awaiting = true;
|
||||
self.marks = 0;
|
||||
self.deadline = Some(now + REANCHOR_FREEZE_MAX);
|
||||
}
|
||||
|
||||
/// Fold one decoded frame and decide whether to present or withhold it.
|
||||
///
|
||||
/// `wire_flags` is the AU's `user_flags` word ([`crate::session::Frame::flags`] /
|
||||
/// `PunktfunkFrame.flags`); the gate reads [`FLAG_SOF`](crate::packet::FLAG_SOF) (the host sets it
|
||||
/// only on IDR AUs — the codec-agnostic keyframe signal the platform decoders don't expose),
|
||||
/// [`USER_FLAG_RECOVERY_ANCHOR`] and [`USER_FLAG_RECOVERY_POINT`]. `decoder_keyframe` is an optional
|
||||
/// belt from decoders that flag IDRs themselves (libavcodec's `AV_FRAME_FLAG_KEY` on Linux/Windows);
|
||||
/// pass `false` where the decoder doesn't (Android MediaCodec, Apple VideoToolbox) and rely on the
|
||||
/// wire `FLAG_SOF`.
|
||||
///
|
||||
/// A decoded frame always clears the no-output streak. When frozen, a live mark stream pushes the
|
||||
/// backstop out ([`RECOVERY_MARK_PATIENCE`]) so a healing wave isn't pre-empted by a mid-heal IDR.
|
||||
///
|
||||
/// [`USER_FLAG_RECOVERY_ANCHOR`]: crate::packet::USER_FLAG_RECOVERY_ANCHOR
|
||||
/// [`USER_FLAG_RECOVERY_POINT`]: crate::packet::USER_FLAG_RECOVERY_POINT
|
||||
pub fn on_decoded(&mut self, wire_flags: u32, decoder_keyframe: bool, now: Instant) -> GateVerdict {
|
||||
self.no_output_streak = 0;
|
||||
let is_keyframe = decoder_keyframe || (wire_flags & FLAG_SOF as u32 != 0);
|
||||
let has_anchor = wire_flags & USER_FLAG_RECOVERY_ANCHOR != 0;
|
||||
let has_mark = wire_flags & USER_FLAG_RECOVERY_POINT != 0;
|
||||
if has_mark && self.awaiting {
|
||||
self.deadline = Some(now + RECOVERY_MARK_PATIENCE);
|
||||
}
|
||||
let (lift, marks) = reanchor_after_frame(is_keyframe, has_anchor, has_mark, self.marks);
|
||||
self.marks = marks;
|
||||
if lift {
|
||||
self.awaiting = false;
|
||||
self.deadline = None;
|
||||
}
|
||||
if self.awaiting {
|
||||
GateVerdict::Hold
|
||||
} else {
|
||||
GateVerdict::Present
|
||||
}
|
||||
}
|
||||
|
||||
/// A received AU produced no decoded frame (decode error, or the decoder swallowed a
|
||||
/// reference-missing delta). Returns `true` when the streak has tripped and the client should
|
||||
/// (throttled) request a keyframe — arming the freeze at the same time, since the stream is broken
|
||||
/// regardless of whether the throttle lets the request through this iteration.
|
||||
pub fn on_no_output(&mut self, now: Instant) -> bool {
|
||||
self.no_output_streak += 1;
|
||||
if self.no_output_streak >= NO_OUTPUT_KEYFRAME_STREAK {
|
||||
self.arm(now);
|
||||
self.no_output_streak = 0;
|
||||
true
|
||||
} else {
|
||||
false
|
||||
}
|
||||
}
|
||||
|
||||
/// Periodic fold of the session's `frames_dropped` counter plus the overdue backstop. Returns
|
||||
/// `true` when the client should (throttled) request a keyframe: either the drop count climbed (a
|
||||
/// fresh unrecoverable loss — arm the freeze) or the freeze has held a full [`REANCHOR_FREEZE_MAX`]
|
||||
/// window with no re-anchor (re-ask and keep holding — NEVER resume to the concealed picture; a
|
||||
/// genuinely dead stream is the QUIC idle-timeout watchdog's job, not the gate's).
|
||||
pub fn poll(&mut self, frames_dropped: u64, now: Instant) -> bool {
|
||||
let mut want_keyframe = false;
|
||||
if frames_dropped > self.last_dropped {
|
||||
self.last_dropped = frames_dropped;
|
||||
self.arm(now);
|
||||
want_keyframe = true;
|
||||
}
|
||||
if self.awaiting && self.deadline.is_some_and(|d| now >= d) {
|
||||
self.deadline = Some(now + REANCHOR_FREEZE_MAX);
|
||||
want_keyframe = true;
|
||||
}
|
||||
want_keyframe
|
||||
}
|
||||
|
||||
/// Whether the gate is currently withholding concealed frames (frozen on the last good picture).
|
||||
pub fn is_holding(&self) -> bool {
|
||||
self.awaiting
|
||||
}
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
// Simulate the gate's re-anchor state across a sequence of decoded frames: each `(is_keyframe,
|
||||
// has_mark)` pair is folded through `reanchor_after_frame`, returning the frame index (0-based) at
|
||||
// which the freeze first lifts, or `None` if it never does. A reset to 0 models a fresh loss
|
||||
// re-arming the freeze (the gate zeroes the count at every arm site).
|
||||
fn lift_at(frames: &[(bool, bool)]) -> Option<usize> {
|
||||
let mut marks = 0u32;
|
||||
for (i, &(is_kf, has_mark)) in frames.iter().enumerate() {
|
||||
// The intra-refresh-mark model never carries an LTR-RFI anchor (that path is exercised by
|
||||
// `an_rfi_anchor_lifts_immediately`), so `has_anchor` is always false here.
|
||||
let (lift, m) = reanchor_after_frame(is_kf, false, has_mark, marks);
|
||||
marks = m;
|
||||
if lift {
|
||||
return Some(i);
|
||||
}
|
||||
}
|
||||
None
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_single_recovery_mark_does_not_lift() {
|
||||
// The first wave boundary after a loss is only half-healed — one mark must hold the freeze.
|
||||
assert_eq!(REANCHOR_MARKS_TO_LIFT, 2);
|
||||
assert_eq!(lift_at(&[(false, true)]), None);
|
||||
assert_eq!(lift_at(&[(false, false), (false, true), (false, false)]), None);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn the_second_recovery_mark_lifts() {
|
||||
// Two marks = a full wave swept after the loss → clean re-anchor.
|
||||
assert_eq!(lift_at(&[(false, true), (false, true)]), Some(1));
|
||||
assert_eq!(
|
||||
lift_at(&[(false, false), (false, true), (false, false), (false, true)]),
|
||||
Some(3)
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_real_keyframe_lifts_immediately() {
|
||||
// An IDR is always a clean anchor — no marks needed.
|
||||
assert_eq!(lift_at(&[(true, false)]), Some(0));
|
||||
assert_eq!(lift_at(&[(false, true), (true, false)]), Some(1));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_fresh_gap_resets_the_mark_count() {
|
||||
// The gate zeroes `marks` at each arm site, so one mark before a new gap plus one after must
|
||||
// NOT lift — the model resets the running count to imitate that.
|
||||
let mut marks = 0u32;
|
||||
let (_, m) = reanchor_after_frame(false, false, true, marks); // mark #1 (pre-gap)
|
||||
marks = m;
|
||||
assert_eq!(marks, 1);
|
||||
marks = 0; // a new gap re-arms the freeze → count reset
|
||||
let (lift, m) = reanchor_after_frame(false, false, true, marks); // first mark of the new wave
|
||||
assert!(!lift, "a single post-gap mark must not lift");
|
||||
assert_eq!(m, 1);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn an_rfi_anchor_lifts_immediately() {
|
||||
// An LTR-RFI recovery anchor is a WHOLE re-anchor (a clean P-frame off a known-good reference),
|
||||
// so — like an IDR — it lifts on the FIRST occurrence, no two-mark wait.
|
||||
let (lift, marks) = reanchor_after_frame(false, true, false, 0);
|
||||
assert!(lift, "an RFI anchor must lift the freeze immediately");
|
||||
assert_eq!(marks, 0, "a lift resets the running mark count");
|
||||
// Even with zero prior marks and no keyframe, the anchor alone is sufficient.
|
||||
let (lift, _) = reanchor_after_frame(false, true, true, 1);
|
||||
assert!(lift, "an anchor lifts regardless of the pending mark count");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn contiguous_indices_are_not_a_gap() {
|
||||
assert_eq!(index_gap(5, 5), None);
|
||||
assert_eq!(index_gap(0, 0), None);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_forward_jump_reports_the_skip_count() {
|
||||
assert_eq!(index_gap(5, 6), Some(1)); // one frame missing
|
||||
assert_eq!(index_gap(5, 9), Some(4));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_straggler_behind_us_is_not_a_gap() {
|
||||
// The reassembler emitted a newer frame first; the late one must not re-arm.
|
||||
assert_eq!(index_gap(9, 5), None);
|
||||
assert_eq!(index_gap(1, 0), None);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn the_index_counter_wraps_cleanly() {
|
||||
// last frame = u32::MAX, so the next expected wraps to 0.
|
||||
assert_eq!(index_gap(0, 0), None);
|
||||
// waiting on u32::MAX, frame 0 arrived → MAX was skipped.
|
||||
assert_eq!(index_gap(u32::MAX, 0), Some(1));
|
||||
assert_eq!(index_gap(u32::MAX, 2), Some(3));
|
||||
// an old frame arriving just after the wrap is still a straggler.
|
||||
assert_eq!(index_gap(0, u32::MAX), None);
|
||||
}
|
||||
|
||||
// ---- gate-level sequence tests (the whole behavioural contract) ----
|
||||
|
||||
const SOF: u32 = FLAG_SOF as u32; // IDR wire flag
|
||||
const ANCHOR: u32 = USER_FLAG_RECOVERY_ANCHOR;
|
||||
const POINT: u32 = USER_FLAG_RECOVERY_POINT;
|
||||
|
||||
fn t0() -> Instant {
|
||||
Instant::now()
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_clean_link_never_holds() {
|
||||
// Disarmed gate presents every frame, keyframe or not, and never asks for anything.
|
||||
let mut g = ReanchorGate::new(0);
|
||||
let now = t0();
|
||||
assert_eq!(g.on_decoded(0, false, now), GateVerdict::Present);
|
||||
assert_eq!(g.on_decoded(SOF, true, now), GateVerdict::Present);
|
||||
assert!(!g.is_holding());
|
||||
assert!(!g.poll(0, now));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_gap_holds_until_the_wire_keyframe_lifts() {
|
||||
// Android/Apple path: no decoder keyframe flag, lift comes from the wire FLAG_SOF alone.
|
||||
let mut g = ReanchorGate::new(0);
|
||||
let now = t0();
|
||||
g.arm(now); // frame-index gap
|
||||
assert!(g.is_holding());
|
||||
assert_eq!(g.on_decoded(0, false, now), GateVerdict::Hold); // concealed delta withheld
|
||||
assert_eq!(g.on_decoded(0, false, now), GateVerdict::Hold);
|
||||
assert_eq!(g.on_decoded(SOF, false, now), GateVerdict::Present); // IDR re-anchors
|
||||
assert!(!g.is_holding());
|
||||
assert_eq!(g.on_decoded(0, false, now), GateVerdict::Present); // stays presenting
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_gap_lifts_on_the_first_rfi_anchor() {
|
||||
let mut g = ReanchorGate::new(0);
|
||||
let now = t0();
|
||||
g.arm(now);
|
||||
assert_eq!(g.on_decoded(0, false, now), GateVerdict::Hold);
|
||||
assert_eq!(g.on_decoded(ANCHOR, false, now), GateVerdict::Present);
|
||||
assert!(!g.is_holding());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_gap_lifts_on_the_second_recovery_mark() {
|
||||
let mut g = ReanchorGate::new(0);
|
||||
let now = t0();
|
||||
g.arm(now);
|
||||
assert_eq!(g.on_decoded(POINT, false, now), GateVerdict::Hold); // first boundary: half-healed
|
||||
assert_eq!(g.on_decoded(0, false, now), GateVerdict::Hold);
|
||||
assert_eq!(g.on_decoded(POINT, false, now), GateVerdict::Present); // second: clean
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_second_gap_mid_freeze_resets_the_marks() {
|
||||
let mut g = ReanchorGate::new(0);
|
||||
let now = t0();
|
||||
g.arm(now);
|
||||
assert_eq!(g.on_decoded(POINT, false, now), GateVerdict::Hold); // mark #1
|
||||
g.arm(now); // a fresh loss re-arms → mark count zeroed
|
||||
assert_eq!(g.on_decoded(POINT, false, now), GateVerdict::Hold); // this is mark #1 of the new wave
|
||||
assert_eq!(g.on_decoded(POINT, false, now), GateVerdict::Present); // #2 lifts
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn the_dropped_climb_arms_and_asks() {
|
||||
let mut g = ReanchorGate::new(5);
|
||||
let now = t0();
|
||||
assert!(!g.poll(5, now), "no climb → no ask"); // baseline
|
||||
assert!(g.poll(6, now), "a climb asks for a keyframe");
|
||||
assert!(g.is_holding(), "and arms the freeze");
|
||||
assert!(!g.poll(6, now), "same value → no repeat ask from the drop path");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn the_no_output_streak_trips_at_three() {
|
||||
let mut g = ReanchorGate::new(0);
|
||||
let now = t0();
|
||||
assert!(!g.on_no_output(now));
|
||||
assert!(!g.on_no_output(now));
|
||||
assert!(g.on_no_output(now), "third no-output trips the streak");
|
||||
assert!(g.is_holding());
|
||||
// A decoded frame resets the streak.
|
||||
g.on_decoded(SOF, false, now); // lifts + resets streak
|
||||
assert!(!g.on_no_output(now));
|
||||
assert!(!g.on_no_output(now));
|
||||
assert!(g.on_no_output(now));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn an_overdue_freeze_re_asks_but_keeps_holding() {
|
||||
let mut g = ReanchorGate::new(0);
|
||||
let start = t0();
|
||||
g.arm(start);
|
||||
// Before the deadline: holding, no re-ask.
|
||||
assert!(!g.poll(0, start));
|
||||
assert!(g.is_holding());
|
||||
// Past REANCHOR_FREEZE_MAX with no re-anchor: re-ask, still holding.
|
||||
let later = start + REANCHOR_FREEZE_MAX + Duration::from_millis(1);
|
||||
assert!(g.poll(0, later), "overdue freeze re-asks for a keyframe");
|
||||
assert!(g.is_holding(), "but never resumes to the concealed picture");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_live_mark_stream_pushes_the_deadline_out() {
|
||||
// A healing wave (marks arriving) must not be pre-empted by the overdue IDR floor.
|
||||
let mut g = ReanchorGate::new(0);
|
||||
let start = t0();
|
||||
g.arm(start);
|
||||
// A mark past the original freeze deadline pushes it out by RECOVERY_MARK_PATIENCE.
|
||||
let t = start + REANCHOR_FREEZE_MAX + Duration::from_millis(10);
|
||||
assert_eq!(g.on_decoded(POINT, false, t), GateVerdict::Hold); // mark #1, deadline pushed
|
||||
// At a time that WOULD have been overdue on the original deadline, poll does not re-ask.
|
||||
assert!(!g.poll(0, t + Duration::from_millis(1)));
|
||||
assert!(g.is_holding());
|
||||
}
|
||||
}
|
||||
@@ -25,7 +25,10 @@
|
||||
// TTL of a v2 envelope; `punktfunk_connection_next_rumble` is unchanged and drops it). Additive —
|
||||
// the wire is backward-compatible (the envelope is a length-tolerant tail on 0xCA), so
|
||||
// [`WIRE_VERSION`] is unchanged.
|
||||
#define ABI_VERSION 5
|
||||
// v6: added the `punktfunk_reanchor_gate_*` surface (post-loss freeze-until-reanchor gate for the
|
||||
// Swift client; Rust embedders use [`reanchor::ReanchorGate`] directly). Additive, client-local —
|
||||
// no wire change, so [`WIRE_VERSION`] is unchanged.
|
||||
#define ABI_VERSION 6
|
||||
|
||||
// The punktfunk/1 **wire** version — what `Hello`/`Welcome` carry and hosts equality-check.
|
||||
// Deliberately its own constant: [`ABI_VERSION`] tracks the embeddable **C surface**
|
||||
@@ -586,6 +589,23 @@
|
||||
#define ColorInfo_MC_BT2020_NCL 9
|
||||
#endif
|
||||
|
||||
// Consecutive no-output AUs that force a keyframe request. ~50 ms at 60 Hz — long enough not to fire
|
||||
// on a one-frame decoder hiccup, short enough that a lost initial IDR (or a mid-GOP join) unfreezes
|
||||
// almost immediately instead of never.
|
||||
#define NO_OUTPUT_KEYFRAME_STREAK 3
|
||||
|
||||
// How many host intra-refresh recovery marks ([`USER_FLAG_RECOVERY_POINT`]) must arrive since the
|
||||
// latest loss before the gate lifts its freeze on an IDR-free stream. TWO, not one: with a continuous
|
||||
// rolling wave the host marks phase-fixed wave boundaries, so the FIRST boundary after a loss is only
|
||||
// partially healed — stripes swept BEFORE the loss still reference the lost frame — and lifting there
|
||||
// would flash a partially-stale picture. The SECOND boundary guarantees a full wave swept entirely
|
||||
// after the loss, so the picture is clean. This stays correct under repeated loss because every fresh
|
||||
// arm resets the count. The cost is up to ~2 wave periods of holding the last good frame — the
|
||||
// deliberate "hold longer, never show garbage" trade.
|
||||
//
|
||||
// [`USER_FLAG_RECOVERY_POINT`]: crate::packet::USER_FLAG_RECOVERY_POINT
|
||||
#define REANCHOR_MARKS_TO_LIFT 2
|
||||
|
||||
// Stable C ABI status codes. `Ok` is 0; all errors are negative so callers can
|
||||
// test `rc < 0`. Do not renumber existing variants — only append.
|
||||
enum PunktfunkStatus
|
||||
@@ -714,6 +734,18 @@ typedef struct PunktfunkConnection PunktfunkConnection;
|
||||
// Opaque session handle. Pointer-only from C.
|
||||
typedef struct PunktfunkSession PunktfunkSession;
|
||||
|
||||
// The shared post-loss freeze state machine. A client feeds it three kinds of event — an *arm* (a
|
||||
// loss was detected: a frame-index gap, a dropped-count climb, or a decoder wedge/demotion), each
|
||||
// *decoded frame* ([`on_decoded`](Self::on_decoded), which decides present-vs-hold and interprets the
|
||||
// re-anchor wire flags), and each *no-output* AU ([`on_no_output`](Self::on_no_output)) — plus a
|
||||
// periodic [`poll`](Self::poll) that folds the dropped counter and fires the overdue backstop.
|
||||
//
|
||||
// The gate emits *intents* only: [`on_no_output`](Self::on_no_output) and [`poll`](Self::poll) return
|
||||
// `true` when the client should ask the host for a keyframe. The client routes that through its own
|
||||
// ~100 ms request throttle (and the precise RFI-vs-keyframe range decision stays in the loss-range
|
||||
// tracker behind [`crate::client::NativeClient::note_frame_index`]) — the gate never touches the wire.
|
||||
typedef struct ReanchorGate ReanchorGate;
|
||||
|
||||
// Forward-compatible session configuration. The caller MUST set `struct_size` to
|
||||
// `sizeof(PunktfunkConfig)`; the core uses it to detect ABI skew.
|
||||
typedef struct {
|
||||
@@ -1737,6 +1769,63 @@ void punktfunk_connection_disconnect_quit(PunktfunkConnection *c);
|
||||
void punktfunk_connection_close(PunktfunkConnection *c);
|
||||
#endif
|
||||
|
||||
// Create a re-anchor gate seeded with the session's current `frames_dropped` (so the first
|
||||
// [`punktfunk_reanchor_gate_poll`] doesn't read the baseline as a loss). Free with
|
||||
// [`punktfunk_reanchor_gate_free`]. Never returns NULL.
|
||||
ReanchorGate *punktfunk_reanchor_gate_new(uint64_t frames_dropped);
|
||||
|
||||
// Free a gate created by [`punktfunk_reanchor_gate_new`]. NULL is a no-op.
|
||||
//
|
||||
// # Safety
|
||||
// `g` was returned by [`punktfunk_reanchor_gate_new`] and is not used after this call.
|
||||
void punktfunk_reanchor_gate_free(ReanchorGate *g);
|
||||
|
||||
// Arm the freeze: a loss was detected (a frame-index gap, or a decoder wedge/demotion). Zeroes the
|
||||
// recovery-mark count and (re-)sets the backstop deadline. NULL is a no-op.
|
||||
//
|
||||
// # Safety
|
||||
// `g` is a valid gate handle.
|
||||
void punktfunk_reanchor_gate_arm(ReanchorGate *g);
|
||||
|
||||
// Fold one decoded frame and write to `out_present` whether to display it (`true`) or withhold it as
|
||||
// a post-loss concealment (`false`). `flags` is the AU's `user_flags` word ([`PunktfunkFrame::flags`]):
|
||||
// the gate reads `FLAG_SOF` (the host's IDR marker), `USER_FLAG_RECOVERY_ANCHOR` and
|
||||
// `USER_FLAG_RECOVERY_POINT`. Pass `decoder_keyframe = false` where the platform decoder doesn't flag
|
||||
// IDRs (VideoToolbox/MediaCodec) — the wire `FLAG_SOF` covers it.
|
||||
//
|
||||
// # Safety
|
||||
// `g` is a valid gate handle; `out_present` is writable or NULL.
|
||||
PunktfunkStatus punktfunk_reanchor_gate_on_decoded(ReanchorGate *g,
|
||||
uint32_t flags,
|
||||
bool decoder_keyframe,
|
||||
bool *out_present);
|
||||
|
||||
// A received AU produced no decoded frame. Writes to `out_request_kf` whether the no-output streak has
|
||||
// tripped and the client should (throttled) request a keyframe — the gate arms the freeze at the same
|
||||
// time.
|
||||
//
|
||||
// # Safety
|
||||
// `g` is a valid gate handle; `out_request_kf` is writable or NULL.
|
||||
PunktfunkStatus punktfunk_reanchor_gate_on_no_output(ReanchorGate *g,
|
||||
bool *out_request_kf);
|
||||
|
||||
// Periodic fold of the session's `frames_dropped` counter plus the overdue backstop. Writes to
|
||||
// `out_request_kf` whether the client should (throttled) request a keyframe (a drop-count climb armed
|
||||
// a fresh freeze, or the freeze is overdue and re-asks while it keeps holding).
|
||||
//
|
||||
// # Safety
|
||||
// `g` is a valid gate handle; `out_request_kf` is writable or NULL.
|
||||
PunktfunkStatus punktfunk_reanchor_gate_poll(ReanchorGate *g,
|
||||
uint64_t frames_dropped,
|
||||
bool *out_request_kf);
|
||||
|
||||
// Whether the gate is currently withholding concealed frames (frozen on the last good picture).
|
||||
// Writes `false` on a NULL gate.
|
||||
//
|
||||
// # Safety
|
||||
// `g` is a valid gate handle; `out_holding` is writable or NULL.
|
||||
PunktfunkStatus punktfunk_reanchor_gate_is_holding(const ReanchorGate *g, bool *out_holding);
|
||||
|
||||
#ifdef __cplusplus
|
||||
} // extern "C"
|
||||
#endif // __cplusplus
|
||||
|
||||
Reference in New Issue
Block a user