https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119625
--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
I think
--- gcc/lto-opts.cc.jj 2025-01-02 11:23:09.939608681 +0100
+++ gcc/lto-opts.cc 2025-04-04 15:02:17.881941293 +0200
@@ -96,17 +96,17 @@ lto_write_options (void)
if (!OPTION_SET_P (flag_cf_protection))
{
- append_to_collect_gcc_options (
- &temporary_obstack, &first_p,
- global_options.x_flag_cf_protection == CF_NONE
- ? "-fcf-protection=none"
- : global_options.x_flag_cf_protection == CF_FULL
- ? "-fcf-protection=full"
- : global_options.x_flag_cf_protection == CF_BRANCH
- ? "-fcf-protection=branch"
- : global_options.x_flag_cf_protection == CF_RETURN
- ? "-fcf-protection=return"
- : "");
+ const char *cf_protection = NULL;
+ switch (global_options.x_flag_cf_protection)
+ {
+ case CF_NONE: cf_protection = "-fcf-protection=none"; break;
+ case CF_FULL: cf_protection = "-fcf-protection=full"; break;
+ case CF_BRANCH: cf_protection = "-fcf-protection=branch"; break;
+ case CF_RETURN: cf_protection = "-fcf-protection=return"; break;
+ }
+ if (cf_protection)
+ append_to_collect_gcc_options (&temporary_obstack, &first_p,
+ cf_protection);
}
/* If debug info is enabled append -g. */
would fix this but whether it fixes also the propagation of "" into the @ file,
I have no idea. With -fhardened flag_cf_protection is CF_FULL | CF_SET,
CF_FULL set by -fhardened handling and CF_SET by the i386 backend:
opts->x_flag_cf_protection
= (cf_protection_level) (opts->x_flag_cf_protection | CF_SET);
Now, whether lto-opts.cc should also switch
(global_options.x_flag_cf_protection & ~CF_SET), I have no idea.
Note,
append_to_collect_gcc_options (&temporary_obstack, &first_p,
global_options.x_flag_pic == 2
? "-fPIC"
: global_options.x_flag_pic == 1
? "-fpic"
: global_options.x_flag_pie == 2
? "-fPIE"
: global_options.x_flag_pie == 1
? "-fpie"
: "-fno-pie");
is also weird. Because I get -fPIC even when I haven't specified that, I've
used -fhardened which implies in this case -fPIE. But for -fPIE, both flag_pic
and flag_pie are equal to 2. So IMHO we should test for flag_pie first before
flag_pic.
