https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98125
--- Comment #17 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
As a quick hack, we could do e.g.
--- gcc/configure.ac 2021-03-23 19:42:05.417907561 +0100
+++ gcc/configure.ac 2021-03-30 14:54:31.655766205 +0200
@@ -3404,12 +3404,15 @@ AC_DEFINE_UNQUOTED(HAVE_GAS_SHF_GNU_RETA
# Test if the assembler supports the section flag 'o' for specifying
# section with link-order.
case "${target}" in
- # Solaris may use GNU assembler with Solairs ld. Even if GNU
+ # Solaris may use GNU assembler with Solaris ld. Even if GNU
# assembler supports the section flag 'o', it doesn't mean that
# Solairs ld supports it.
*-*-solaris2*)
gcc_cv_as_section_link_order=no
;;
+ powerpc*-*-*)
+ gcc_cv_as_section_link_order=no
+ ;;
*)
gcc_GAS_CHECK_FEATURE([section 'o' flag], gcc_cv_as_section_link_order,
[2,35,0], [--fatal-warnings],
i.e. effectively revert H.J's patch for powerpc* targets.
That is a hack rather than proper solution.
For GCC11 we need some solution at least for the P1 regression, i.e. the
ELFv1 support that worked in the past and was broken by
r11-5656-g694d4a6d0c466d0fbc97920a9c6641a7b349ca35
