https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116185
Andrew Pinski <pinskia at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |ASSIGNED Ever confirmed|0 |1 Last reconfirmed| |2024-08-03 Assignee|unassigned at gcc dot gnu.org |pinskia at gcc dot gnu.org --- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> --- Note I think this is the same issue as in PR 116189. Basically GCSE hashtable entries is based on the number of instructions which can cause different hashtable walk which then will cause different register #s being used. And it just happened that late_combine does some slight changes that the register allocator then gets a different answer due to the psedu-register numbers being different. So mine for now. How the issue with GCSE hashtables and their walk has not shown up until now is shocking and why it showed up on sh only is another mystery. THis has been a latent bug since we emitting debug insn and making sure we get the same output for with/without -g too.