https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69019
Bug ID: 69019 Summary: [5 regression] Failure of test case gcc.target/powerpc/ti_math1.c Product: gcc Version: 5.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: target Assignee: unassigned at gcc dot gnu.org Reporter: seurer at linux dot vnet.ibm.com Target Milestone: --- PASS: gcc.target/powerpc/ti_math1.c (test for excess errors) PASS: gcc.target/powerpc/ti_math1.c scan-assembler-times addc 1 FAIL: gcc.target/powerpc/ti_math1.c scan-assembler-times adde 1 PASS: gcc.target/powerpc/ti_math1.c scan-assembler-times subfc 1 PASS: gcc.target/powerpc/ti_math1.c scan-assembler-times subfe 1 PASS: gcc.target/powerpc/ti_math1.c scan-assembler-not subf Seen with gcc-5-branch revision 231903. I don't know when this started failing This test works with gcc 4.9 and gcc 6 and works with gcc 5 on big endian. The problem can be seen here when comparing the assembler from gcc 6 with that from 5: seurer@genoa:~/tests/gcc$ diff ti_math1.s.gcc5 ti_math1.s.gcc6 13,14c13 < addze 5,5 < add 4,5,10 --- > adde 4,5,10 It also only occurs with -O2 With -O2: addze 5,5 add 4,5,10 With -O1: adde 4,5,10