https://sourceware.org/bugzilla/show_bug.cgi?id=19985
--- Comment #10 from Mike Frysinger <vapier at gentoo dot org> --- (In reply to Alan Modra from comment #9) counter point: bfd already does it for powerpc linux targets (and more): ppc -> enable BE & LE, 32 & 64, default to BE/32 ppcle -> enable BE & LE, 32 & 64, default to LE/32 ppc64 -> enable BE & LE, 32 & 64, default to BE/64 ppc64le -> enable BE & LE, 32 & 64, default to LE/64 the linker should be consistent, and including these targets by default adds very little overhead. it already has been at least to get biarch support. you're right that gcc doesn't do this by default, but it's trivial to pass --enable-targets=all, and also adds very little overhead. plus, i expect most people do this already in order to just get biarch support (32bit+64bit) which is needed to build a 64bit kernel from a 32bit userland (and a 32bit kernel from a 64bit userland which still makes sense in the VM case). this does not require a full toolchain as it only affects gcc's codegen steps. which is the only thing grub cares about since it's producing bare metal code here. so what's the problem with having the linker be consistent with bfd and including the full 32/64 be/le matrix by default ? afaict, at least bfd has behaved this way for over a decade and no one has noticed/cared/complained. -- You are receiving this mail because: You are on the CC list for the bug. _______________________________________________ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
