https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84877
--- Comment #23 from Hans-Peter Nilsson <hp at gcc dot gnu.org> --- It seems this was fixed by r17-4649-g8aaa5f6e62075c, leaving the test XPASSing. At least that's the case for cris-elf, and judging from posts to gcc-testresults, for hppa (32-bit) too. I can't say for sure for sparc, because reports say it XPASSed before, but that may have been spurious as it doesn't haven't opted in to the stabilization machinery I added to the test-case.
