On 11/13/23 10:30, juzhe.zh...@rivai.ai wrote: > I just checked definition of REG_EQUAL and REG_EQUIV. > > As you said, REG_EQUIV is more reasonable. Agree with use rtx_equal_p on > REG_EQUIV and skip REG_EQUAL. > Could you check whether it does fix your issues ?
Yes it would fix the issues. I just figured we could work a bit harder and also catch cases where two "different" REG_EQUALS would still be the same. But I'm not sure whether such cases exist at all (leaning towards no...). Going to post v2 after the tests ran. Regards Robin