https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123976
--- Comment #236 from Iain Sandoe <iains at gcc dot gnu.org> --- you already tried a different make - make 4.x right the kernel and most of the tools on a G5 powermac are still 32b, rosetta2 makes use of the same kernel code and the same tools and the same SDK. this libgcc makefile recipe has been in use for a long time, and has worked just fine with native powerpc make-3.81 your Makefile and mine are the same (modulo different pathnames) the dependencies are in the make fragments, but those are included the same for both makefiles and the source code from the zip file is the same as the sourcecode for the git checkout I have (I checked). So, yes, it is still possible that there's some previously-unseen failure with powerpc-make-3.81 that has never affected the build before, and does not affect a rosetta build and does not affect a 32b x86 build (after all the source code for the tools is the same for both). it just seems more likely that the problem lies elsewhere.
