https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124266
--- Comment #14 from Richard Sandiford <rsandifo at gcc dot gnu.org> --- (In reply to Richard Biener from comment #13) > Target maintainers morphed this into sth not P1. The ICE is a regression from GCC 16, so I think that is still P1. I think this PR has effectively become two separate bugs: the pair-fusion thing that is the immediate cause of the ICE, and the target bug that causes excess CFA_RESTORE notes.
