Files
punktfunk/.gitea/workflows/ci.yml
T
enricobuehlerandClaude Fable 5 143a707f76 fix(zerocopy): the mechanical sweep batch — truthful fence waits, forgiving env flags, no leaked planes
From the pf-zerocopy review sweep (design/pf-zerocopy-sweep-handoff.md), the
compile-verifiable batch:

- dmabuf_fence: the blocking poll's result was discarded — EINTR silently
  skipped the wait (reopening the stale-frame race a SIGCHLD away) and a
  timeout reported as waited. Now EINTR retries with the remaining budget
  and the caller gets Signaled/TimedOut/NoFence, so the one diagnostic
  operators have about implicit fencing stops lying. (C2)
- env flags: PUNKTFUNK_ZEROCOPY=TRUE meant *off* — values are case-folded
  now, and an unrecognised spelling falls back to the flag's default with a
  one-shot warning instead of silently inverting the operator's intent. (C3)
- cuda: alloc_pitched_nv12 leaked the Y plane when the UV allocation failed
  (per-frame under VRAM pressure — the worst possible time to leak); a failed
  async-copy enqueue now drains the stream before returning, so a recycled
  pool buffer can't race an orphaned in-flight copy. (C4, C5)
- worker: --fd is validated (>= 3, fstat + S_ISSOCK) before OwnedFd adoption
  — 'zerocopy-worker --fd -1' was constructing OwnedFd's niche value. (V1)
- docs: all 15 rustdoc warnings fixed, including the link to the renamed
  note_raw_dmabuf_negotiation_failed. (D1)
- CI: a SPIR-V drift gate — the committed .spv blobs are include_bytes!'d and
  rebuilt by hand; the gate diffs disassembly (filtering only the
  shaderc/glslang generator difference) so a forgotten rebuild fails CI
  instead of shipping the old kernel. Both blobs verified in sync. (S1)
- tests: bt709_limited pinned to external BT.709 anchors (it is the sole
  oracle for the GPU colour self-test); blend_geometry gets its first tests —
  empty rect, CURSOR_MAX clamp, per-format group counts, and negative-ox
  floor alignment. (T2, T3)

Verified on .25: clippy -D warnings clean, 27/27 tests, cargo doc 0 warnings.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-28 12:40:37 +02:00

244 lines
12 KiB
YAML

# CI for punktfunk (Gitea Actions). Linux jobs run on the `ubuntu-latest` runner; the Rust
# job runs inside the prebuilt builder image (ci/rust-ci.Dockerfile — system FFmpeg 8,
# PipeWire, GL/GBM, libcuda link stub, pinned-channel rustup) so the workspace links the
# same libs as the dev boxes. Apple client CI lives in apple.yml (macOS runner).
name: ci
on:
push:
branches: [main]
pull_request:
jobs:
rust:
runs-on: ubuntu-24.04
container:
image: git.unom.io/unom/punktfunk-rust-ci:latest
timeout-minutes: 90
steps:
- uses: actions/checkout@v4
# punktfunk-client-linux link deps. Also baked into rust-ci.Dockerfile — but ci.yml
# runs against the image from the PREVIOUS push (docker.yml bootstrap note), so this
# keeps the job green across image-content changes; a no-op once the image has them.
- name: GTK4/libadwaita/SDL3 dev packages
run: |
apt-get update
apt-get install -y --no-install-recommends libgtk-4-dev libadwaita-1-dev libsdl3-dev
# The committed pf-zerocopy SPIR-V blobs are pulled in with include_bytes! and rebuilt only
# by hand — edit a .comp, forget the rebuild, and the OLD kernel ships with no compile error
# or failing test. Recompile each shader and diff the disassembly. Filtering OpSourceExtension
# (+ --no-header) is exactly what absorbs the shaderc-vs-glslang generator difference; every
# instruction, ID and constant must match.
- name: Shader SPIR-V drift gate (pf-zerocopy)
run: |
apt-get install -y --no-install-recommends glslang-tools spirv-tools
for s in rgb2nv12_buf cursor_blend; do
d=crates/pf-zerocopy/src/imp
glslangValidator -V "$d/$s.comp" -o "/tmp/$s.spv" >/dev/null
diff <(spirv-dis --no-header "$d/$s.spv" | grep -v OpSourceExtension) \
<(spirv-dis --no-header "/tmp/$s.spv" | grep -v OpSourceExtension) \
|| { echo "::error::$d/$s.spv is stale — rebuild it from $s.comp"; exit 1; }
done
# Best-effort caches (act_runner's built-in cache server). Keyed on Cargo.lock:
# registry/git are download caches, target/ the incremental build. The target key
# carries the rustc version — resolved via `rustc --version` (below) rather than parsed
# from rust-toolchain.toml, so a pin bump there invalidates stale incremental state too.
- name: Cache keys
run: echo "rustc=$(rustc --version | cut -d' ' -f2)" >> "$GITHUB_ENV"
- uses: actions/cache@v4
with:
path: |
/usr/local/cargo/registry
/usr/local/cargo/git
key: cargo-home-${{ hashFiles('Cargo.lock') }}
restore-keys: cargo-home-
- uses: actions/cache@v4
with:
path: target
# -v3-: the prior `cargo-target-<rustc>-*` cache was poisoned when the runner ran
# out of disk mid-build and actions/cache saved a truncated target/ (a dep's .rmeta
# went missing -> E0463 "can't find crate"). A suffix bump wouldn't help — restore-keys
# would fall back to the poisoned prefix — so the prefix itself is versioned.
key: cargo-target-v3-${{ env.rustc }}-${{ hashFiles('Cargo.lock') }}
restore-keys: cargo-target-v3-${{ env.rustc }}-
- name: Format
run: cargo fmt --all --check
- name: Clippy (deny warnings)
run: cargo clippy --workspace --all-targets --locked -- -D warnings
- name: Build
run: cargo build --workspace --locked
- name: Test (unit + loopback + proptest + C ABI harness)
run: cargo test --workspace --locked
# The GPU encode backends are OFF by default, so every step above compiles ~none of them:
# `nvenc` gates enc/linux/nvenc_cuda.rs (+ nvenc_core/nvenc_status) and `vulkan-encode` gates
# enc/linux/vulkan_video.rs (+ the vendored vk_av1_encode/vk_valve_rgb bindings) — ~8,150
# lines carrying ~70 `unsafe` blocks. Their ONLY prior CI coverage was deb.yml's
# `cargo build`, where warnings are not errors, so pf-encode's own
# `#![deny(clippy::undocumented_unsafe_blocks)]` — the crate's stated unsafe-proof gate —
# was never actually enforced on them. (`pyrowave` needs no extra step: punktfunk-host has
# `default = ["pyrowave"]`, so the steps above already cover it.)
#
# `--all-targets` is load-bearing, not decoration: without it the feature-gated
# `#[cfg(test)]` modules are never compiled, which is exactly how all ten
# `NvencCudaEncoder::open` call sites in nvenc_cuda.rs's tests drifted to the wrong arity
# (E0061 x10) without any job noticing.
#
# GPU-free: every test needing real hardware is `#[ignore]`d, and NVENC/CUDA resolve their
# entry points at RUNTIME (dlopen), so the test binary links without a driver present.
# (On MSVC the same crate link-imports those symbols instead, which is why windows-host.yml
# can only type-check these tests via clippy — see the note there.)
#
# Scoped to `-p pf-encode` with ITS OWN feature names: punktfunk-host has no code gated on
# `nvenc`/`vulkan-encode` (its only `cfg(feature)` sites are the two `pyrowave` ones in
# capture.rs, and pyrowave is default-on, so the steps above already cover them). Going
# through `--features punktfunk-host/...` would force punktfunk-host into the selection and
# re-run its entire test suite a second time for no extra coverage.
#
# `pyrowave` is listed explicitly even though it is punktfunk-host's default: selecting only
# `-p pf-encode` takes the host out of the resolution, and pf-encode's own default is empty.
# Naming it keeps this the SHIPPED Linux feature set — deb.yml builds
# `--features punktfunk-host/nvenc,punktfunk-host/vulkan-encode` WITHOUT
# `--no-default-features`, so the .deb carries nvenc + vulkan-encode + pyrowave together, and
# that combination is what deserves the lint.
- name: Clippy + test the feature-gated Linux encode backends
run: |
cargo clippy -p pf-encode --all-targets --locked \
--features nvenc,vulkan-encode,pyrowave -- -D warnings
cargo test -p pf-encode --locked --features nvenc,vulkan-encode,pyrowave
- name: C ABI harness (standalone link proof)
run: bash crates/punktfunk-core/tests/c/run.sh
- name: Verify generated header is committed & up to date
run: |
cargo build -p punktfunk-core --locked
git config --global --add safe.directory "$PWD"
git diff --exit-code include/punktfunk_core.h \
|| (echo "include/punktfunk_core.h is stale — commit the regenerated header" && exit 1)
# The client stack cross-checked for aarch64. NOT an artifact job — deb.yml ships those —
# this exists so a portability defect fails here instead of surfacing in a release build or
# on a user's board. It earns its runtime: the bug that motivated it (a Vulkan extension
# array typed `*const i8`, where `c_char` is signed on x86_64 and UNSIGNED on aarch64)
# compiled cleanly on every target CI built at the time.
#
# Client crates only, listed explicitly: the host's encode stack is x86 (NVENC/QSV/AMF) and
# `--workspace` would drag it in. Runs in the cross image (amd64 toolchain + arm64 sysroot,
# ci/rust-ci-arm64cross.Dockerfile) on the ordinary runner — no arm64 runner involved.
rust-arm64:
runs-on: ubuntu-24.04
container:
image: git.unom.io/unom/punktfunk-rust-ci-arm64cross:latest
timeout-minutes: 60
steps:
- uses: actions/checkout@v4
- name: Cache keys
run: echo "rustc=$(rustc --version | cut -d' ' -f2)" >> "$GITHUB_ENV"
- uses: actions/cache@v4
with:
path: |
/usr/local/cargo/registry
/usr/local/cargo/git
key: cargo-home-${{ hashFiles('Cargo.lock') }}
restore-keys: cargo-home-
- uses: actions/cache@v4
with:
path: target
# Its OWN prefix: aarch64 artifacts must never share the amd64 jobs' target cache.
key: cargo-target-arm64-v1-${{ env.rustc }}-${{ hashFiles('Cargo.lock') }}
restore-keys: cargo-target-arm64-v1-${{ env.rustc }}-
- name: Clippy for aarch64 (deny warnings)
run: |
cargo clippy --target aarch64-unknown-linux-gnu --all-targets --locked \
-p punktfunk-core -p pf-client-core -p pf-presenter -p pf-console-ui \
-p punktfunk-client-session -p punktfunk-client-linux \
-- -D warnings
# The minimal embedded build — no Skia, no PyroWave — is what a small image installs, so
# it has to keep compiling on its own, not just as a subset of the default features.
- name: Build the session binary, minimal features
run: |
cargo build --release --target aarch64-unknown-linux-gnu --locked \
-p punktfunk-client-session --no-default-features
web:
runs-on: ubuntu-24.04
container:
image: oven/bun:1
timeout-minutes: 30
defaults:
run:
working-directory: web
steps:
# oven/bun ships neither git nor a real node (only a bun shim) — actions/checkout
# needs both. The slim Debian base also lacks ca-certificates, so without it git's
# HTTPS fetch of the repo dies with "Problem with the SSL CA cert (path? access
# rights?)" — no CA bundle to validate git.unom.io's (public) Let's Encrypt cert.
- name: Install git + node + CA certs
working-directory: /
run: apt-get update && apt-get install -y --no-install-recommends ca-certificates git nodejs
- uses: actions/checkout@v4
- name: Install dependencies
run: bun install --frozen-lockfile --ignore-scripts
# Build first: it generates the orval API client + paraglide messages that
# typechecking imports.
- name: Build
run: bun run build
- name: Typecheck
run: bun run lint
docs-site:
runs-on: ubuntu-24.04
container:
image: oven/bun:1
timeout-minutes: 30
defaults:
run:
working-directory: docs-site
steps:
# ca-certificates: the slim Debian base lacks a CA bundle, so actions/checkout's
# HTTPS fetch otherwise fails with "Problem with the SSL CA cert" (see web job).
- name: Install git + CA certs
working-directory: /
run: apt-get update && apt-get install -y --no-install-recommends ca-certificates git
- uses: actions/checkout@v4
- name: Install dependencies
run: bun install --frozen-lockfile --ignore-scripts
# Build first: fumadocs-mdx emits the .source typegen the typecheck imports.
- name: Build
run: bun run build
- name: Typecheck
run: bun run lint
bench:
# Tier-1 (criterion microbenchmarks) + Tier-2 (FEC loss recovery) — GPU-free, so they run here.
# Report-only: prints the numbers + a diff vs the committed baseline to the job summary and never
# fails the build (shared CI hardware is too noisy to gate on). The tight regression gate + the
# real encode/stream path live on the self-hosted GPU runner (Tier 3, bench-gpu.yml).
runs-on: ubuntu-24.04
container:
image: git.unom.io/unom/punktfunk-rust-ci:latest
timeout-minutes: 30
steps:
- uses: actions/checkout@v4
- name: Prep
run: |
git config --global --add safe.directory "$PWD"
command -v python3 >/dev/null || { apt-get update && apt-get install -y --no-install-recommends python3; }
- name: Tier-1 microbenchmarks (criterion)
run: cargo bench -p punktfunk-core --bench pipeline -- --warm-up-time 1 --measurement-time 3
- name: Tier-2 FEC loss recovery (loss-harness)
run: cargo run -q -p loss-harness
- name: Compare vs baseline (report-only)
run: python3 scripts/bench/compare.py --threshold 0.5