[Kernel] Migrate scattered quantization kernels to sglang.kernels (RFC #29630, Phase 2.5, 1/7) (#30784)

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

test/manual/quant/test_block_fp8.py

@@ -4,16 +4,16 @@ from functools import lru_cache
 
 import torch
 
-from sglang.srt.layers.activation import SiluAndMul
-from sglang.srt.layers.moe.moe_runner.triton_utils.fused_moe import fused_moe
-from sglang.srt.layers.moe.topk import TopKConfig, select_experts
-from sglang.srt.layers.quantization.fp8_kernel import (
+from sglang.kernels.ops.quantization.fp8_kernel import (
     per_tensor_quant_mla_fp8,
     per_token_group_quant_fp8,
     per_token_group_quant_mla_deep_gemm_masked_fp8,
     static_quant_fp8,
     w8a8_block_fp8_matmul,
 )
+from sglang.srt.layers.activation import SiluAndMul
+from sglang.srt.layers.moe.moe_runner.triton_utils.fused_moe import fused_moe
+from sglang.srt.layers.moe.topk import TopKConfig, select_experts
 from sglang.srt.layers.quantization.fp8_utils import (
     input_to_float8,
     mxfp8_group_quantize,

test/manual/quant/test_custom_ops.py

@@ -7,7 +7,7 @@ import sys
 import pytest
 import torch
 
-from sglang.srt.layers.quantization.fp8_kernel import is_fp8_fnuz, scaled_fp8_quant
+from sglang.kernels.ops.quantization.fp8_kernel import is_fp8_fnuz, scaled_fp8_quant
 from sglang.srt.utils import is_cuda, is_hip
 
 _is_cuda = is_cuda()