From e1d45bc280e4eb8954da84dea69e0026ca6ec470 Mon Sep 17 00:00:00 2001
From: kyleliang-nv <kylliang@nvidia.com>
Date: Tue, 16 Sep 2025 17:34:37 -0700
Subject: [PATCH] Fix decord dependency for aarch64 docker build (#10529)

---
 .github/workflows/release-docker-gb200.yml |  2 +-
 docker/Dockerfile.gb200                    | 20 ++++++++++++++++----
 python/pyproject.toml                      |  3 ++-
 3 files changed, 19 insertions(+), 6 deletions(-)

diff --git a/.github/workflows/release-docker-gb200.yml b/.github/workflows/release-docker-gb200.yml
index fbcacb330..87de03b85 100644
--- a/.github/workflows/release-docker-gb200.yml
+++ b/.github/workflows/release-docker-gb200.yml
@@ -33,4 +33,4 @@ jobs:
           version=$(cat python/sglang/version.py | cut -d'"' -f2)
           tag=v${version}-cu129-gb200
 
-          docker buildx build --platform linux/arm64 --push --output type=image -t lmsysorg/sglang:${tag} -f docker/Dockerfile.gb200 --build-arg CUDA_VERSION=12.9.1 --build-arg BUILD_TYPE=blackwell --no-cache .
+          docker buildx build --platform linux/arm64 --push --output type=image -t lmsysorg/sglang:${tag} -f docker/Dockerfile.gb200 --build-arg CUDA_VERSION=12.9.1 --build-arg BUILD_TYPE=blackwell_aarch64 --no-cache .
diff --git a/docker/Dockerfile.gb200 b/docker/Dockerfile.gb200
index c76ae8640..c3a2c5083 100644
--- a/docker/Dockerfile.gb200
+++ b/docker/Dockerfile.gb200
@@ -1,7 +1,8 @@
 ARG CUDA_VERSION=12.9.1
-FROM nvidia/cuda:${CUDA_VERSION}-cudnn-devel-ubuntu22.04
+FROM nvidia/cuda:${CUDA_VERSION}-cudnn-devel-ubuntu22.04 as base
 
-ARG BUILD_TYPE=blackwell
+ARG BUILD_TYPE=blackwell_aarch64
+ARG BRANCH_TYPE=remote
 ARG DEEPEP_COMMIT=1b14ad661c7640137fcfe93cccb2694ede1220b0
 ARG CMAKE_BUILD_PARALLEL_LEVEL=2
 ARG SGL_KERNEL_VERSION=0.3.10
@@ -62,10 +63,21 @@ RUN mkdir -p /tmp/gdrcopy && cd /tmp \
 # Fix DeepEP IBGDA symlink
 RUN ln -sf /usr/lib/$(uname -m)-linux-gnu/libmlx5.so.1 /usr/lib/$(uname -m)-linux-gnu/libmlx5.so
 
-# Clone and install SGLang
+FROM scratch AS local_src
+COPY . /src
+
+FROM base AS build-image
+# Install SGLang
 WORKDIR /sgl-workspace
+ARG BRANCH_TYPE
+COPY --from=local_src /src /tmp/local_src
+RUN if [ "$BRANCH_TYPE" = "local" ]; then \
+        cp -r /tmp/local_src /sgl-workspace/sglang; \
+    else \
+        git clone --depth=1 https://github.com/sgl-project/sglang.git /sgl-workspace/sglang; \
+    fi \
+ && rm -rf /tmp/local_src
 RUN python3 -m pip install --no-cache-dir --upgrade pip setuptools wheel html5lib six \
- && git clone --depth 1 https://github.com/sgl-project/sglang.git \
  && cd sglang \
  && case "$CUDA_VERSION" in \
       12.9.1) CUINDEX=129 ;; \
diff --git a/python/pyproject.toml b/python/pyproject.toml
index d7a1a36ef..5cb35b006 100755
--- a/python/pyproject.toml
+++ b/python/pyproject.toml
@@ -89,7 +89,8 @@ tracing = [
     "opentelemetry-exporter-otlp-proto-grpc",
 ]
 all = ["sglang[test]", "sglang[decord]"]
-blackwell = ["nvidia-cutlass-dsl==4.2.0", "sglang[test]"]
+blackwell = ["nvidia-cutlass-dsl==4.2.0", "sglang[test]", "sglang[decord]"]
+blackwell_aarch64 = ["nvidia-cutlass-dsl==4.2.0", "sglang[test]"]
 dev = ["sglang[test]"]
 
 [project.urls]