how to use docker to build TensorRT OSS 8.2.0 aarch64

Description

see also x.log

-- [download 49% complete]
[ 53%] Linking CXX shared library ../out/libnvinfer_plugin.so
/usr/bin/ld: read in flex scanner failed
clang-12: error: linker command failed with exit code 1 (use -v to see invocation)
make[2]: *** [plugin/CMakeFiles/nvinfer_plugin.dir/build.make:2219: out/libnvinfer_plugin.so.8.2.1] Error 1
make[2]: Leaving directory '/root/TensorRT/TensorRT-8.2.1/build'
make[1]: *** [CMakeFiles/Makefile2:1500: plugin/CMakeFiles/nvinfer_plugin.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....

x.log

Environment

TensorRT Version: 8.2.5.1
NVIDIA GPU: A100
NVIDIA Driver Version: 470.141.03
CUDA Version: 11.4
CUDNN Version: 8.2
Baremetal or Container (if so, version): none

steps

export TRT_LIBPATH=/root/TensorRT/TensorRT-8.2.5.1/lib/
export TRT_OSSPATH=/root/TensorRT/TensorRT-8.2.1
cmake .. -DTRT_LIB_DIR=$TRT_LIBPATH -DTRT_OUT_DIR=`pwd`/out -DTRT_PLATFORM_ID=aarch64 -DCUDA_VERSION=11.4 -Dnvinfer_LIB_PATH=$TRT_LIBPATH -Dnvparsers_LIB_PATH=$TRT_LIBPATH

Do I have to use docker to build a container to solve this problem? If so, will the 18.09 version of docker work?