Skip to content

feat(docker): enable HW-codec features in Dockerfile.full-gpu #321

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
streamer45 merged 1 commit into from
Apr 18, 2026
Merged

feat(docker): enable HW-codec features in Dockerfile.full-gpu #321

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 27 additions & 3 deletions Dockerfile.full-gpu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,13 +13,19 @@ ARG SHERPA_ONNX_VERSION=1.12.17
# This is the recommended version for ONNX Runtime GPU in 2025
ARG CUDA_VERSION=12.3.2

# Stage 0: Extract CUDA dev headers for nvcodec build
# shiguredo_nvcodec uses bindgen to generate FFI bindings and needs cuda.h.
ARG CUDA_VERSION
FROM nvidia/cuda:${CUDA_VERSION}-devel-ubuntu22.04 AS cuda-dev

# Stage 1: Build Rust dependencies
FROM rust:1.92-slim-bookworm AS rust-deps

WORKDIR /build

# Install build dependencies (includes all deps needed by workspace plugins)
# meson + ninja-build + python3 are needed by dav1d_static (builds dav1d from source)
# libva-dev + libgbm-dev are needed by cros-codecs (vaapi feature)
RUN apt-get update && apt-get install -y \
pkg-config \
libssl-dev \
Expand All @@ -35,8 +41,13 @@ RUN apt-get update && apt-get install -y \
meson \
ninja-build \
python3 \
libva-dev \
libgbm-dev \
&& rm -rf /var/lib/apt/lists/*

# Copy CUDA headers for nvcodec feature (shiguredo_nvcodec needs cuda.h)
COPY --from=cuda-dev /usr/local/cuda/include /usr/local/cuda/include

# Copy workspace files to build dependencies (plugins are now excluded)
COPY Cargo.toml Cargo.lock ./
COPY apps ./apps
Expand All @@ -60,7 +71,8 @@ RUN --mount=type=cache,target=/usr/local/cargo/registry \
--mount=type=cache,target=/usr/local/cargo/git \
--mount=type=cache,target=/build/target \
RUSTFLAGS="-C link-arg=-fuse-ld=bfd" \
cargo build --profile release-lto -p streamkit-server --bin skit --features "moq,svt_av1_static,dav1d_static" -j$(nproc) && \
CUDA_INCLUDE_PATH=/usr/local/cuda/include \
cargo build --profile release-lto -p streamkit-server --bin skit --features "moq,svt_av1_static,dav1d_static,nvcodec,vulkan_video,vaapi" -j$(nproc) && \
Copy link
Copy Markdown
Contributor Author

@staging-devin-ai-integration staging-devin-ai-integration Bot Apr 17, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🚩 No Vulkan SDK dev package installed for vulkan_video build

The vulkan_video feature depends on the vk-video crate, but neither Stage 1 nor Stage 3 installs libvulkan-dev (the Vulkan development package). This is likely fine because Rust Vulkan crates (typically built on ash) use runtime dynamic loading (dlopen) for libvulkan.so.1 rather than link-time linking against libvulkan.so. However, if vk-video ever changes to use link-time binding, the build would fail. The runtime stage correctly installs libvulkan1 at Dockerfile.full-gpu:695.

Staging: Open in Devin

Was this helpful? React with 👍 or 👎 to provide feedback.

Debug

Playground

Copy link
Copy Markdown
Contributor Author

@staging-devin-ai-integration staging-devin-ai-integration Bot Apr 17, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Correct — vk-video uses ash which does runtime dlopen("libvulkan.so.1"), so no link-time dependency on libvulkan-dev. The CI E2E workflow (e2e.yml) also builds with vulkan_video without installing libvulkan-dev and it works fine. The runtime stage already has libvulkan1 at line 695.

# Copy compiled artifacts out of cache mount so they persist in the layer
mkdir -p /build/target-out && \
cp -r /build/target/release-lto /build/target-out/
Expand Down Expand Up @@ -100,8 +112,13 @@ RUN apt-get update && apt-get install -y \
meson \
ninja-build \
python3 \
libva-dev \
libgbm-dev \
&& rm -rf /var/lib/apt/lists/*

# Copy CUDA headers for nvcodec feature (shiguredo_nvcodec needs cuda.h)
COPY --from=cuda-dev /usr/local/cuda/include /usr/local/cuda/include

# Copy workspace files (plugins are excluded from workspace)
COPY Cargo.toml Cargo.lock ./
COPY apps ./apps
Expand Down Expand Up @@ -133,7 +150,8 @@ RUN --mount=type=cache,target=/usr/local/cargo/registry \
# Build only the server binary \
# (-fuse-ld=bfd: use system ld instead of rust-lld — see note above) \
RUSTFLAGS="-C link-arg=-fuse-ld=bfd" \
cargo build --profile release-lto --features "moq,svt_av1_static,dav1d_static" --bin skit -j$(nproc); \
CUDA_INCLUDE_PATH=/usr/local/cuda/include \
cargo build --profile release-lto --features "moq,svt_av1_static,dav1d_static,nvcodec,vulkan_video,vaapi" --bin skit -j$(nproc); \
# Copy final binary out of cache mount \
mkdir -p /build/bin && cp /build/target/release-lto/skit /build/bin/skit \
'
Expand Down Expand Up @@ -660,7 +678,9 @@ ARG CUDA_VERSION
FROM nvidia/cuda:${CUDA_VERSION}-cudnn9-runtime-ubuntu22.04

# Install runtime dependencies and debugging tools
# libvulkan1: required by wgpu compositor for GPU-accelerated compositing via Vulkan
# libvulkan1: required by wgpu compositor and vulkan_video HW codec via Vulkan
# libva2 + libva-drm2 + libgbm1: VA-API runtime for vaapi HW codec
# mesa-va-drivers: VA-API backends for Intel/AMD GPUs (NVIDIA uses nvidia-vaapi-driver from host)
# libfdk-aac2 is in Ubuntu's multiverse component.
RUN sed -i 's/main restricted/main restricted multiverse/' /etc/apt/sources.list && \
apt-get update && apt-get install -y \
Expand All @@ -673,6 +693,10 @@ RUN sed -i 's/main restricted/main restricted multiverse/' /etc/apt/sources.list
fonts-dejavu-core \
libfdk-aac2 \
libvulkan1 \
libva2 \
libva-drm2 \
libgbm1 \
mesa-va-drivers \
gdb \
curl \
iproute2 \
Expand Down
Loading