[Intel GPU] DeepSeek V4 9/N: use sgl-kernel implementation of hadamard_transform on XPU (#28046)

Co-authored-by: Rahul Vijayaraghavan <rahul.vijayaraghavan@intel.com>
Co-authored-by: Ma Mingfei <mingfei.ma@intel.com>

python/sglang/srt/layers/attention/dsa/dsa_indexer.py

@@ -49,6 +49,7 @@ from sglang.srt.utils import (
     is_gfx95_supported,
     is_hip,
     is_npu,
+    is_xpu,
 )
 from sglang.srt.utils.custom_op import register_custom_op
 
@@ -58,6 +59,7 @@ global _use_multi_stream
 _is_cuda = is_cuda()
 _is_hip = is_hip()
 _is_npu = is_npu()
+_is_xpu = is_xpu()
 
 if not _is_npu:
     from sglang.jit_kernel.dsa import (
@@ -338,6 +340,8 @@ def rotate_activation(x: torch.Tensor) -> torch.Tensor:
     # from sgl_kernel import hadamard_transform
     if _is_hip:
         from fast_hadamard_transform import hadamard_transform
+    elif _is_xpu:
+        from sgl_kernel import hadamard_transform
     else:
         from sglang.jit_kernel.hadamard import hadamard_transform