Wire enable_cuda_graph_decode through engine args, stage config and CLI

Follows the vae_use_tiling path: OmniDiffusionConfig field, OmniEngineArgs,
StageDeployConfig, _DiffusionConfigProjection and --[no-]enable-cuda-graph-decode.
Documents interplay with enforce_eager and cudagraph_mode.

vllm_omni/diffusion/data.py

@@ -955,6 +955,18 @@ class OmniDiffusionConfig:
 
     # Compilation
     enforce_eager: bool = False
+    # Capture fixed-shape KV-cache decode (denoising) steps into CUDA graphs.
+    # Currently only Qwen-Image-2.1's transformer implements this; other models
+    # ignore the flag. This is a per-model diffusion path, independent of the
+    # AR engine's ``compilation_config.cudagraph_mode`` (which does not apply
+    # to diffusion stages) and of ``diffusion_compile_granularity`` — models
+    # with decode graphs skip automatic torch.compile of their DiT blocks, so
+    # the two never stack. ``enforce_eager=True`` forces eager decode and
+    # disables capture regardless of this flag; unsupported configurations
+    # (SP/TP, ring, HSDP, dynamic LoRA, padded masks, quantized prefix KV, or
+    # a second in-flight request aliasing the same graph key) log and fall
+    # back to eager decode.
+    enable_cuda_graph_decode: bool = True
     # Controls the generic compilation path used when a pipeline does not
     # provide its own setup_compile() implementation.
     diffusion_compile_granularity: str = "regional"