Re: libstdc++ support for m68k

2016-09-12 Thread Sebastian Huber
It seems that the libstdc++ configure script already detects that for RTEMS the generic atomic built-ins are available and doesn't select the CPU-specific header file. So, this stuff in libstdc++-v3/config/cpu/m68k/atomicity.h looks like dead code. On 09/09/16 21:54, Joel Sherrill wrote: I am

libstdc++ support for m68k

2016-09-09 Thread Sebastian Huber
Hello, I was a bit surprised to find this special case for RTEMS in libstdc++ on m68k: https://gcc.gnu.org/ml/libstdc++/2016-09/msg00017.html Any objections to eliminate this special case and simply use the generic implementation? We have to check if the atomic operations for our m68k (or