https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113030
--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> --- Or rather there is a typo in the script. This fixes the script's chkarch and chkcpu commands: ``` diff --git a/gcc/config/arm/parsecpu.awk b/gcc/config/arm/parsecpu.awk index b827d71dded..fdc6e35a7ce 100644 --- a/gcc/config/arm/parsecpu.awk +++ b/gcc/config/arm/parsecpu.awk @@ -529,7 +529,7 @@ function check_cpu (name) { for (n = 2; n <= exts; n++) { if (!((cpu_name, extensions[n]) in cpu_opt_remove) \ - && !((cpu_name, extensions[n]) in cpu_optaliases)) { + && !((cpu_name, extensions[n]) in cpu_opt_alias)) { return "error" } } @@ -552,7 +552,7 @@ function check_arch (name) { for (n = 2; n <= exts; n++) { if (!((extensions[1], extensions[n]) in arch_opt_remove) \ - && !((extensions[1], extensions[n]) in arch_optaliases)) { + && !((extensions[1], extensions[n]) in arch_opt_alias)) { return "error" } } ``` But it does not change how -march=/-mcpu= is handled inside cc1.