On 31 May 2016 at 11:35, Samuel Thibault <sthiba...@debian.org> wrote: > So reassigning to cuda, perhaps a proper fix can be done in the .h files > there without having to ask upstream :)
I'll have to check if Nvidia's license allows us to fix bugs in their headers. ;) Here we have a minimal reproducer from the Ubuntu bug [1]: $ echo '#include <stdio.h>' > dummy.cpp $ g++ -c -O2 dummy.cpp $ cp dummy.cpp dummy.cu $ nvcc -c -O2 dummy.cu /usr/include/string.h: In function ‘void* __mempcpy_inline(void*, const void*, size_t)’: /usr/include/string.h:652:42: error: ‘memcpy’ was not declared in this scope return (char *) memcpy (__dest, __src, __n) + __n; Again, defining _FORCE_INLINES works around the problem: $ nvcc -c -O2 -D_FORCE_INLINES dummy.cu [1] https://bugs.launchpad.net/ubuntu/+source/nvidia-cuda-toolkit/+bug/1589751