On Nov 25 2019, Bernd Schmidt wrote: > On 11/25/19 12:26 PM, Andreas Schwab wrote: >> On Nov 24 2019, Bernd Schmidt wrote: >> >>> Whew, I think I have it. One tst instruction eliminated when it >>> shouldn't have been: >>> >>> move.w %a4,%d0 >>> - tst.b %d0 >>> - jeq .L352 >>> + jeq .L353 >>> >>> And the reason - that's a movqi using move.w. The following should fix >>> it. >> >> Apparently that also fixed the testsuite regressions. > > I still wish I knew how you managed to reproduce those.
It's undefined behaviour, so you can get all kinds of random results. The binary diff clearly shows the missing tstb insns. Andreas. -- Andreas Schwab, SUSE Labs, sch...@suse.de GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7 "And now for something completely different."
