[Bug rtl-optimization/68506] [6 Regression] wrong code at -O3 on x86_64-linux-gnu

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68506

--- Comment #4 from ktkachov at gcc dot gnu.org ---
I'm testing a fix, but there's a slight issue with the testcase.
__assert_fail () doesn't exist for non-glibc libraries i.e. newlib so the
testcase fails to link there.
I tried replacing the __assert_fail with __builtin_abort but the bug doesn't
trigger then. Any ideas on how to make this portable?
If not, do we have any dejagnu directives that check for a glibc target?