https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64159
Richard Biener <rguenth at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P3 |P1 Component|tree-optimization |target Summary|FAIL: |[5 Regression] FAIL: |gcc.dg/tree-ssa/ssa-dom-cse |gcc.dg/tree-ssa/ssa-dom-cse |-2.c scan-tree-dump |-2.c scan-tree-dump |optimized "return 28;" |optimized "return 28;" --- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> --- It's already xfailed for hppa, so yes, likely (if you can verify the theory). Note that there may be other reasons, like vectorizing using stmts that we don't constant-fold or the inability to unroll the loop due to target specific limits. Needs investigation. It's still a testsuite regression.