[BugFix] Fix missing symbol triggering FA2 fallback on Hopper

[LucasWilkinson]

fix (missing symbol introduced in: #105)

>>> import vllm.vllm_flash_attn._vllm_fa3_C
/home/mgoin/code/vllm/.venv/lib64/python3.12/site-packages/torch/cuda/__init__.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
  import pynvml  # type: ignore[import]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: /home/mgoin/code/vllm/vllm/vllm_flash_attn/_vllm_fa3_C.abi3.so: undefined symbol: _Z12run_mha_fwd_ILi80EN7cutlass10bfloat16_tELi128ELi128ELb1ELb1ELb0ELb1EEvR16Flash_fwd_paramsP11CUstream_st