I try to cross compile host computer and Jetson Xavier, made by matlab coder

I try to compile c++ code generated by matlab coder on Jetson Xavier, ROS, using cmake.

I success generating c++ code without ploblem.

but I got a ploblem when compilingg on Xavier. I think error occured beacause of cross compile.

when I generate c++ code by matlab coder, select option nvidia jetson and GNU linux64bit and static library (.a)

I need your help

Thanks.

below is error code

/usr/bin/ld: /home/nvidia/catkin_ws/src/realtime_inpainting/src/wrapper/wrapper.a(wrapper.o): Relocations in generic ELF (EM: 62)

/usr/bin/ld: /home/nvidia/catkin_ws/src/realtime_inpainting/src/wrapper/wrapper.a(wrapper.o): Relocations in generic ELF (EM: 62)

/usr/bin/ld: /home/nvidia/catkin_ws/src/realtime_inpainting/src/wrapper/wrapper.a(wrapper.o): Relocations in generic ELF (EM: 62)

/usr/bin/ld: /home/nvidia/catkin_ws/src/realtime_inpainting/src/wrapper/wrapper.a(wrapper.o): Relocations in generic ELF (EM: 62)

/usr/bin/ld: /home/nvidia/catkin_ws/src/realtime_inpainting/src/wrapper/wrapper.a(wrapper.o): Relocations in generic ELF (EM: 62)

/usr/bin/ld: /home/nvidia/catkin_ws/src/realtime_inpainting/src/wrapper/wrapper.a(wrapper.o): Relocations in generic ELF (EM: 62)

/home/nvidia/catkin_ws/src/realtime_inpainting/src/wrapper/wrapper.a: error adding symbols: File in wrong format

collect2: error: ld returned 1 exit status

realtime_inpainting/CMakeFiles/inpainting_cuda.dir/build.make:131: recipe for target '/home/nvidia/catkin_ws/devel/lib/realtime_inpainting/inpainting_cuda' failed

make[2]: *** [/home/nvidia/catkin_ws/devel/lib/realtime_inpainting/inpainting_cuda] Error 1

CMakeFiles/Makefile2:7310: recipe for target 'realtime_inpainting/CMakeFiles/inpainting_cuda.dir/all' failed

make[1]: *** [realtime_inpainting/CMakeFiles/inpainting_cuda.dir/all] Error 2

make[1]: *** Waiting for unfinished jobs....

[David Fink]

Based on some forums with the same error, it seems like the compiler and linker are building for different architectures, so the linker isn't very happy with the object files the compiler produced.

If you're building for Jetson, don't you need to use nvcc as the linker instead of usr/bin/ld?

https://devblogs.nvidia.com/separate-compilation-linking-cuda-device-code/

[Ramakrishna Mandalapu]

Hi DaHoon,

Could you please give some more info on the settings you did in the model? I am little confused about the option GNU linux64bit, where did you set this option?

Thanks,

Ramakrishna