ref(upstream): FULL TREE — Deep-Spark xllm (1470) + ds_vllm csrc/models (703)

Replaces cherry-picked upstream_ref with complete source trees.

xllm/ — Iluvatar official C++ inference engine (15MB, 1470 files)
  Complete: kernels → layers → models → runtime → scheduler → api
  Excluded: .git, binary images, third_party submodule checkouts

ds_vllm/ — Iluvatar official vllm fork (8MB, 703 files)
  Included: csrc/ (ALL CUDA kernels), fused_moe/, qwen3_5 model, _custom_ops
  Excluded: tests, benchmarks, docs, examples (not needed for reference)

Critical call chains now fully traceable:
  MoE: moe_topk_softmax_kernels.cuh → ixformer.h → fused_moe.cpp → layer
  GDN: qwen3_gated_delta_net_base.cpp → qwen3_5_gated_delta_net.cpp
  Attention: ixformer.h → xllm_paged_attention → attention.cpp

upstream_ref/xllm/cibuild/build_cuda.sh

@@ -0,0 +1,29 @@
+#!/bin/bash
+set -e
+
+function error() {
+  echo "Require build command, e.g. python setup.py build --device cuda"
+  exit 1
+}
+
+IMAGE="quay.io/jd_xllm/xllm-ai:xllm-dev-cuda-x86"
+
+RUN_OPTS=(
+  --rm
+  -t
+  --privileged
+  --ipc=host
+  --network=host
+  --pid=host
+  --shm-size '128gb'
+  -v /export/home:/export/home
+  -v /export/home/cuda_vcpkg_cache:/root/.cache/vcpkg # cached vcpkg installed dir
+  -w /export/home
+)
+
+CMD="$*"
+[[ -z "${CMD}" ]] && error
+
+[[ ! -x $(command -v docker) ]] && echo "ERROR: 'docker' command is missing." && exit 1
+
+docker run "${RUN_OPTS[@]}" "${IMAGE}" bash -c "set -euo pipefail; cd $(pwd); ${CMD}"