undefined reference to `pthread_create', `pthread_detach`, `pthread_cancel`, `pthread_join`

[jiapei100]

Description

Failed to build with the following error messages:

Linking CXX executable cmTC_a4f90
/usr/bin/cmake -E cmake_link_script CMakeFiles/cmTC_a4f90.dir/link.txt --verbose=1
/usr/local/bin/g++ -Wno-deprecated-declarations  -DBUILD_SYSTEM=cmake_oss  CMakeFiles/cmTC_a4f90.dir/src.cxx.o -o cmTC_a4f90 
/usr/bin/ld: CMakeFiles/cmTC_a4f90.dir/src.cxx.o: in function `main':
src.cxx:(.text+0x2d): undefined reference to `pthread_create'
/usr/bin/ld: src.cxx:(.text+0x39): undefined reference to `pthread_detach'
/usr/bin/ld: src.cxx:(.text+0x45): undefined reference to `pthread_cancel'
/usr/bin/ld: src.cxx:(.text+0x56): undefined reference to `pthread_join'
collect2: error: ld returned 1 exit status
make[1]: *** [CMakeFiles/cmTC_a4f90.dir/build.make:99: cmTC_a4f90] Error 1
make[1]: Leaving directory '....../TensorRT/build/CMakeFiles/CMakeTmp'
make: *** [Makefile:127: cmTC_a4f90/fast] Error 2

Environment

TensorRT Version: 8.0.1
NVIDIA GPU: GeForce RTX 2080 Ti
NVIDIA Driver Version: NVIDIA-SMI 460.84
CUDA Version: 11.3
CUDNN Version: 8.2.1
Operating System: Ubuntu 20.04
Python Version (if applicable): 3.8.10
Tensorflow Version (if applicable): 2.7.0
PyTorch Version (if applicable): 1.10.0
Baremetal or Container (if so, version): N/A

Relevant Files

....../TensorRT/build/CMakeFiles/CMakeTmp

Steps To Reproduce

Just build the source code via git clone.

[bmfire1]

build you project with -pthread

[jiapei100]

@bmfire1

Already linked with pthread by default. Still problematic.

[ttyio]

Hello @jiapei100 , have you tried using container?
https://github.com/NVIDIA/TensorRT/tree/release/8.0#setting-up-the-build-environment

Thanks

[tanndx17]

I had the same error, a little different

Linking CXX executable cmTC_ebefe
/usr/local/bin/cmake -E cmake_link_script CMakeFiles/cmTC_ebefe.dir/link.txt --verbose=1
/usr/bin/g++ -Wno-deprecated-declarations -DBUILD_SYSTEM=cmake_oss -DCMAKE_HAVE_LIBC_PTHREAD
CMakeFiles/cmTC_ebefe.dir/src.cxx.o -o cmTC_ebefe
CMakeFiles/cmTC_ebefe.dir/src.cxx.o: In function main': src.cxx:(.text+0x3e): undefined reference to pthread_create'
src.cxx:(.text+0x4a): undefined reference to pthread_detach' src.cxx:(.text+0x56): undefined reference to pthread_cancel'
src.cxx:(.text+0x67): undefined reference to pthread_join' src.cxx:(.text+0x7b): undefined reference to pthread_atfork'
collect2: error: ld returned 1 exit status
CMakeFiles/cmTC_ebefe.dir/build.make:106: recipe for target 'cmTC_ebefe' failed
make[1]: *** [cmTC_ebefe] Error 1
make[1]: Leaving directory '/data/TensorRT-8.0.1/build/CMakeFiles/CMakeTmp'
Makefile:141: recipe for target 'cmTC_ebefe/fast' failed
make: *** [cmTC_ebefe/fast] Error 2

How did you solved this?

[cherry-jam]

I had the same error too, using container

Linking CXX executable cmTC_10c48
/usr/local/bin/cmake -E cmake_link_script CMakeFiles/cmTC_10c48.dir/link.txt --verbose=1
/usr/bin/g++ -Wno-deprecated-declarations -DBUILD_SYSTEM=cmake_oss CMakeFiles/cmTC_10c48.dir/CheckSymbolExists.cxx.o -o cmTC_10c48
/usr/bin/ld: CMakeFiles/cmTC_10c48.dir/CheckSymbolExists.cxx.o: in function main': CheckSymbolExists.cxx:(.text+0x1f): undefined reference to pthread_create'
collect2: error: ld returned 1 exit status
make[1]: *** [CMakeFiles/cmTC_10c48.dir/build.make:87: cmTC_10c48] Error 1
make[1]: Leaving directory '/workspace/TensorRT-release-8.0/build/CMakeFiles/CMakeTmp'
make: *** [Makefile:121: cmTC_10c48/fast] Error 2

any news about this issue?

[styler00dollar]

I had a similar error.

git clone -b master https://github.com/nvidia/TensorRT TensorRT
cd TensorRT
git submodule update --init --recursive

mkdir -p build && cd build
export TRT_LIBPATH=/content/TensorRT-8.2.3.0
cmake .. -DTRT_LIB_DIR=$TRT_LIBPATH -DTRT_OUT_DIR=`pwd`/out
make -j$(nproc)

Run Build Command(s):/usr/bin/make -f Makefile cmTC_e9692/fast && /usr/bin/make  -f CMakeFiles/cmTC_e9692.dir/build.make CMakeFiles/cmTC_e9692.dir/build
make[1]: Entering directory '/content/TensorRT/build/CMakeFiles/CMakeTmp'
Building CXX object CMakeFiles/cmTC_e9692.dir/src.cxx.o
/usr/bin/g++ -DCMAKE_HAVE_LIBC_PTHREAD   -o CMakeFiles/cmTC_e9692.dir/src.cxx.o -c /content/TensorRT/build/CMakeFiles/CMakeTmp/src.cxx
Linking CXX executable cmTC_e9692
/usr/local/bin/cmake -E cmake_link_script CMakeFiles/cmTC_e9692.dir/link.txt --verbose=1
/usr/bin/g++ CMakeFiles/cmTC_e9692.dir/src.cxx.o -o cmTC_e9692 
CMakeFiles/cmTC_e9692.dir/src.cxx.o: In function `main':
src.cxx:(.text+0x3e): undefined reference to `pthread_create'
src.cxx:(.text+0x4a): undefined reference to `pthread_detach'
src.cxx:(.text+0x56): undefined reference to `pthread_cancel'
src.cxx:(.text+0x67): undefined reference to `pthread_join'
src.cxx:(.text+0x7b): undefined reference to `pthread_atfork'
collect2: error: ld returned 1 exit status
CMakeFiles/cmTC_e9692.dir/build.make:98: recipe for target 'cmTC_e9692' failed
make[1]: *** [cmTC_e9692] Error 1
make[1]: Leaving directory '/content/TensorRT/build/CMakeFiles/CMakeTmp'
Makefile:127: recipe for target 'cmTC_e9692/fast' failed
make: *** [cmTC_e9692/fast] Error 2

I solved it by replacing the variable with a path

cmake .. -DTRT_LIB_DIR=/content/TensorRT-8.2.3.0 -DTRT_OUT_DIR=/content/out

Edit:
It does actually compile a little and does not crash during cmake with that command, but it does crash later during compile. To fix that, use:

cmake .. -DTRT_LIB_DIR=/content/TensorRT-8.2.3.0/lib -DTRT_INC_DIR=/content/TensorRT-8.2.3.0/include -DTRT_OUT_DIR=/content/out

like this github issue suggested.

[jiapei100]

Compatible with protobuf-3.19.4 ?
Can we ignore caffe ?

[ 56%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/caffeParser.cpp.o
cd ....../TensorRT/build/parsers/caffe && /usr/local/bin/g++ -DGOOGLE_PROTOBUF_ARCH_64_BIT -Dgoogle=google_private -I/usr/local/cuda/include -I/include -I....../TensorRT/include -I....../TensorRT/parsers/caffe/. -I....../TensorRT/parsers/caffe/caffeParser -I....../TensorRT/parsers/caffe/caffeParser/opParsers -I....../TensorRT/parsers/caffe/caffeWeightFactory -I....../TensorRT/parsers/caffe/../common -I....../TensorRT/build/third_party.protobuf/include -I....../TensorRT/build/parsers/caffe/proto -Wno-deprecated-declarations  -DBUILD_SYSTEM=cmake_oss -O3 -DNDEBUG -std=c++11 -MD -MT parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/caffeParser.cpp.o -MF CMakeFiles/nvcaffeparser_static.dir/caffeParser/caffeParser.cpp.o.d -o CMakeFiles/nvcaffeparser_static.dir/caffeParser/caffeParser.cpp.o -c ....../TensorRT/parsers/caffe/caffeParser/caffeParser.cpp
In file included from ....../TensorRT/parsers/caffe/caffeParser/caffeParser.cpp:23:
....../TensorRT/parsers/caffe/caffeParser/readProto.h: In function ‘bool nvcaffeparser1::readBinaryProto(trtcaffe::NetParameter*, const char*, size_t)’:
....../TensorRT/parsers/caffe/caffeParser/readProto.h:51:51: error: no matching function for call to ‘google_private::protobuf::io::CodedInputStream::SetTotalBytesLimit(int, int)’
   51 |     codedInput.SetTotalBytesLimit(int(bufSize), -1);
      |                                                   ^
In file included from ....../TensorRT/build/parsers/caffe/proto/trtcaffe.pb.h:23,
                 from ....../TensorRT/parsers/caffe/caffeWeightFactory/caffeWeightFactory.h:26,
                 from ....../TensorRT/parsers/caffe/caffeParser/caffeParser.h:26,
                 from ....../TensorRT/parsers/caffe/caffeParser/caffeParser.cpp:20:
....../TensorRT/build/third_party.protobuf/include/google/protobuf/io/coded_stream.h:406:8: note: candidate: ‘void google_private::protobuf::io::CodedInputStream::SetTotalBytesLimit(int)’
  406 |   void SetTotalBytesLimit(int total_bytes_limit);
      |        ^~~~~~~~~~~~~~~~~~
....../TensorRT/build/third_party.protobuf/include/google/protobuf/io/coded_stream.h:406:8: note:   candidate expects 1 argument, 2 provided
....../TensorRT/parsers/caffe/caffeParser/caffeParser.cpp: In member function ‘virtual const nvcaffeparser1::IBlobNameToTensor* nvcaffeparser1::CaffeParser::parseBuffers(const uint8_t*, std::size_t, const uint8_t*, std::size_t, nvinfer1::INetworkDefinition&, nvinfer1::DataType)’:
....../TensorRT/parsers/caffe/caffeParser/caffeParser.cpp:330:60: error: no matching function for call to ‘google_private::protobuf::io::CodedInputStream::SetTotalBytesLimit(std::size_t&, int)’
  330 |         codedModelStream.SetTotalBytesLimit(modelLength, -1);
      |                                                            ^
In file included from ....../TensorRT/build/parsers/caffe/proto/trtcaffe.pb.h:23,
                 from ....../TensorRT/parsers/caffe/caffeWeightFactory/caffeWeightFactory.h:26,
                 from ....../TensorRT/parsers/caffe/caffeParser/caffeParser.h:26,
                 from ....../TensorRT/parsers/caffe/caffeParser/caffeParser.cpp:20:
....../TensorRT/build/third_party.protobuf/include/google/protobuf/io/coded_stream.h:406:8: note: candidate: ‘void google_private::protobuf::io::CodedInputStream::SetTotalBytesLimit(int)’
  406 |   void SetTotalBytesLimit(int total_bytes_limit);
      |        ^~~~~~~~~~~~~~~~~~
....../TensorRT/build/third_party.protobuf/include/google/protobuf/io/coded_stream.h:406:8: note:   candidate expects 1 argument, 2 provided
....../TensorRT/parsers/caffe/caffeParser/caffeParser.cpp: In member function ‘virtual nvcaffeparser1::IBinaryProtoBlob* nvcaffeparser1::CaffeParser::parseBinaryProto(const char*)’:
....../TensorRT/parsers/caffe/caffeParser/caffeParser.cpp:627:46: error: no matching function for call to ‘google_private::protobuf::io::CodedInputStream::SetTotalBytesLimit(int, int)’
  627 |     codedInput.SetTotalBytesLimit(INT_MAX, -1);
      |                                              ^
In file included from ....../TensorRT/build/parsers/caffe/proto/trtcaffe.pb.h:23,
                 from ....../TensorRT/parsers/caffe/caffeWeightFactory/caffeWeightFactory.h:26,
                 from ....../TensorRT/parsers/caffe/caffeParser/caffeParser.h:26,
                 from ....../TensorRT/parsers/caffe/caffeParser/caffeParser.cpp:20:
....../TensorRT/build/third_party.protobuf/include/google/protobuf/io/coded_stream.h:406:8: note: candidate: ‘void google_private::protobuf::io::CodedInputStream::SetTotalBytesLimit(int)’
  406 |   void SetTotalBytesLimit(int total_bytes_limit);
      |        ^~~~~~~~~~~~~~~~~~
....../TensorRT/build/third_party.protobuf/include/google/protobuf/io/coded_stream.h:406:8: note:   candidate expects 1 argument, 2 provided
make[2]: *** [parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/build.make:438: parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/caffeParser.cpp.o] Error 1
make[2]: Leaving directory '....../TensorRT/build'
make[1]: *** [CMakeFiles/Makefile2:1399: parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/all] Error 2
make[1]: Leaving directory '....../TensorRT/build'
make: *** [Makefile:159: all] Error 2

[nvpohanh]

@jiapei100 Did you compile this within the docker image we provide in README?

[nvpohanh]

Closing due to >14 days without activity. Please feel free to reopen if the issue still exists. Thanks