https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91635
--- Comment #18 from Jim Wilson <wilson at gcc dot gnu.org> --- I was willing to write a patch, I just needed a day to catch up as I'm hopelessly overloaded with work. But since Jakub wrote one, it looks right to me, and I think we should use that. I'm not sure exactly what testing Kito did, and he has now left on vacation, so I will just do my own testing. I think the reason why RISC-V needs paradoxical_reg and other architectures don't is that most architectures have sense enough to define zero and sign extend instructions. RISC-V will finally get that when the draft B (bit manipulation) extension is formally approved.