undefined symbol: eglCreateContext

[ljiqy]

Hi，thank you for your great work！
when I run tf.load_op_library('dirt/librasterise.so'),it has error:
Traceback (most recent call last):
File "", line 1, in
File "/home/lj/software/anaconda3/envs/MGN_py2.7/lib/python2.7/site-packages/tensorflow/python/framework/load_library.py", line 60, in load_op_library
lib_handle = py_tf.TF_LoadLibrary(library_filename)
tensorflow.python.framework.errors_impl.NotFoundError: ../dirt/librasterise.so: undefined symbol: eglCreateContext

and I use ldd -r dirt/librasterise.so to show:
(MGN_py2.7) [root@runonce-lj-syt build]# ldd -r ../dirt/librasterise.so
linux-vdso.so.1 => (0x00007ffe49d93000)
libEGL_nvidia.so.0 => /usr/lib64/libEGL_nvidia.so.0 (0x00007f092a8ba000)
libGLX_nvidia.so.0 => /usr/lib64/libGLX_nvidia.so.0 (0x00007f092a590000)
libtensorflow_framework.so => /home/lj/software/anaconda3/envs/MGN_py2.7/lib/python2.7/site-packages/tensorflow/libtensorflow_framework.so (0x00007f0929622000)
librt.so.1 => /usr/lib64/librt.so.1 (0x00007f092941a000)
libpthread.so.0 => /usr/lib64/libpthread.so.0 (0x00007f09291fe000)
libdl.so.2 => /usr/lib64/libdl.so.2 (0x00007f0928ffa000)
libstdc++.so.6 => /usr/lib64/libstdc++.so.6 (0x00007f092afa1000)
libm.so.6 => /usr/lib64/libm.so.6 (0x00007f0928cf8000)
libgcc_s.so.1 => /usr/lib64/libgcc_s.so.1 (0x00007f0928ae2000)
libc.so.6 => /usr/lib64/libc.so.6 (0x00007f0928720000)
/lib64/ld-linux-x86-64.so.2 (0x00007f092aef9000)
libnvidia-glsi.so.460.32.03 => /usr/lib64/libnvidia-glsi.so.460.32.03 (0x00007f092847e000)
libnvidia-tls.so.460.32.03 => /usr/lib64/libnvidia-tls.so.460.32.03 (0x00007f0928279000)
libnvidia-glcore.so.460.32.03 => /usr/lib64/libnvidia-glcore.so.460.32.03 (0x00007f0926372000)
libX11.so.6 => /usr/lib64/libX11.so.6 (0x00007f0926034000)
libXext.so.6 => /usr/lib64/libXext.so.6 (0x00007f0925e22000)
libcublas.so.9.0 => /usr/local/cuda-9.0/lib64/libcublas.so.9.0 (0x00007f09229ec000)
libcuda.so.1 => /usr/lib64/libcuda.so.1 (0x00007f0921304000)
libcudnn.so.7 => /home/lj/software/cudnn_7.3/lib64/libcudnn.so.7 (0x00007f090f41f000)
libcufft.so.9.0 => /usr/local/cuda-9.0/lib64/libcufft.so.9.0 (0x00007f090737e000)
libcurand.so.9.0 => /usr/local/cuda-9.0/lib64/libcurand.so.9.0 (0x00007f090341a000)
libcudart.so.9.0 => /usr/local/cuda-9.0/lib64/libcudart.so.9.0 (0x00007f09031ad000)
libxcb.so.1 => /usr/lib64/libxcb.so.1 (0x00007f0902f8b000)
libXau.so.6 => /usr/lib64/libXau.so.6 (0x00007f0902d87000)
undefined symbol: eglCreateContext (../dirt/librasterise.so)
undefined symbol: eglGetProcAddress (../dirt/librasterise.so)
undefined symbol: eglMakeCurrent (../dirt/librasterise.so)
undefined symbol: eglGetError (../dirt/librasterise.so)
undefined symbol: eglChooseConfig (../dirt/librasterise.so)
undefined symbol: eglBindAPI (../dirt/librasterise.so)
undefined symbol: eglInitialize (../dirt/librasterise.so)

how can I solve these problem?

[pmh47]

There should be a libEGL linked, not just libEGL_nvidia. Did you have to override any paths/etc. in the build? It should pick up libEGL automatically, and that then calls into libEGL_nvidia

[dev2021-ctrl]

I am facing similar issue but getting this error

tf.load_op_library('dirt/librasterise.so')
Traceback (most recent call last):
File "", line 1, in
File "/home/ubuntu/anaconda3/envs/mgn4/lib/python3.7/site-packages/tensorflow/python/framework/load_library.py", line 61, in load_op_library
lib_handle = py_tf.TF_LoadLibrary(library_filename)
tensorflow.python.framework.errors_impl.NotFoundError: /usr/lib/x86_64-linux-gnu/libOpenGL.so.0: undefined symbol: _glapi_tls_Current

@pmh47 please advise

[ljiqy]

Thank you for your quick reply!
In my system it can't find EGL library automatically, so I edit CMakeCache.txt.
Now I change EGL_LIBRARIES:FILEPATH=/usr/lib64/libEGL.so.1, it can be build successfully, but when I ran python tests/square_test.py, another error shows:
2021-09-06 16:28:41.482959: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 AVX512F FMA
2021-09-06 16:28:41.774868: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1432] Found device 0 with properties:
name: Tesla V100-PCIE-32GB major: 7 minor: 0 memoryClockRate(GHz): 1.38
pciBusID: 0000:63:00.0
totalMemory: 31.75GiB freeMemory: 31.45GiB
2021-09-06 16:28:41.994806: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1432] Found device 1 with properties:
name: Tesla V100-PCIE-32GB major: 7 minor: 0 memoryClockRate(GHz): 1.38
pciBusID: 0000:64:00.0
totalMemory: 31.75GiB freeMemory: 29.94GiB
2021-09-06 16:28:42.000071: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1511] Adding visible gpu devices: 0, 1
2021-09-06 16:28:42.880030: I tensorflow/core/common_runtime/gpu/gpu_device.cc:982] Device interconnect StreamExecutor with strength 1 edge matrix:
2021-09-06 16:28:42.880124: I tensorflow/core/common_runtime/gpu/gpu_device.cc:988] 0 1
2021-09-06 16:28:42.880138: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1001] 0: N Y
2021-09-06 16:28:42.880146: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1001] 1: Y N
2021-09-06 16:28:42.880328: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 30511 MB memory) -> physical GPU (device: 0, name: Tesla V100-PCIE-32GB, pci bus id: 0000:63:00.0, compute capability: 7.0)
2021-09-06 16:28:42.880730: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 29049 MB memory) -> physical GPU (device: 1, name: Tesla V100-PCIE-32GB, pci bus id: 0000:64:00.0, compute capability: 7.0)
2021-09-06 16:28:43.528035: F /home/lj/code/3D/clothing/software/dirt/csrc/gl_common.h:46] extensions eglQueryDevicesEXT, eglQueryDeviceAttribEXT and eglGetPlatformDisplayEXT not available

[MilesTheProwler]

@ljiqy did you solve it ? I have same error like you

2021-09-06 16:28:43.528035: F /home/lj/code/3D/clothing/software/dirt/csrc/gl_common.h:46] extensions eglQueryDevicesEXT, eglQueryDeviceAttribEXT and eglGetPlatformDisplayEXT not available