fix(host): forward reconfigure_bitrate through TrackedEncoder + probe --rebitrate validator

The gpu-session TrackedEncoder wrapper delegates every Encoder method by hand, so the new reconfigure_bitrate fell through to the trait's false default and EVERY bitrate change silently took the rebuild+IDR path — the live .21 A/B caught it (host log said 'rebuilt', never 'in place'). Also: - punktfunk-probe --rebitrate KBPS:SECS — headless mid-stream SetBitrate validator (cursor-wiggles so a damage-driven idle desktop keeps publishing frames through the switch). Live-verified on .21: one NVENC session open, then 'encoder bitrate reconfigured in place (adaptive bitrate — no IDR)' at 20→60 Mbps. - on-hardware nvenc_cuda reconfigure smoke test (20→60→10 Mbps in place, zero IDRs — green on the RTX 5070 Ti). - BitrateChanged doc no longer claims the switch costs an IDR. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-14 20:06:56 +02:00
parent a1af916e38
commit 9bf72cdfb5
3 changed files with 78 additions and 5 deletions
@@ -415,9 +415,10 @@ pub struct SetBitrate {
 }

 /// `host → client`: answer to [`SetBitrate`] — the bitrate the host actually configured (the
-/// request clamped to its supported band). The encoder switches on the next frame (an IDR); the
-/// stream never pauses. Also the controller's liveness signal: no answer ⇒ an old host that
-/// doesn't renegotiate bitrate.
+/// request clamped to its supported band). The encoder retargets in place where the backend can
+/// (no IDR — the stream carries straight on); a backend without in-place reconfigure rebuilds and
+/// switches on the next frame (an IDR). The stream never pauses either way. Also the controller's
+/// liveness signal: no answer ⇒ an old host that doesn't renegotiate bitrate.
 #[derive(Clone, Copy, Debug, PartialEq, Eq)]
 pub struct BitrateChanged {
    pub bitrate_kbps: u32,