https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115042
--- Comment #10 from Vladimir Makarov <vmakarov at gcc dot gnu.org> --- The 1st patch resulted in failures of equivalence substitution in the postponed insns and resulted in wrong code generation on some secondary targets. The 2nd patch variant change postponing algorithm and solves the equivalence problem of the previous variant. The patch also constrains insn postponing only to asm insns. It makes the patch more safe and constrains different code generation too very rare cases.
