https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64460
--- Comment #8 from Joel Sherrill <joel at gcc dot gnu.org> --- (In reply to ktkachov from comment #7) > I guess the testcase is flaky. > I've posted the patch to fix the ICE and the reasoning behind it at > https://gcc.gnu.org/ml/gcc-patches/2015-01/msg00686.html Can you try again with the full test case? The head still fails for me at -O2 and -Os but not -O1. The location in toplev.c has moved down a few lines. FWIW my original report had -mcpu=xscale but it fails with -mtune=xscale as well. I $ ~/test-gcc/install-head/bin/arm-rtems4.11-gcc --version arm-rtems4.11-gcc (GCC) 5.0.0 20150112 (experimental) $ ~/test-gcc/install-head/bin/arm-rtems4.11-gcc -mtune=xscale -O2 /tmp/gumstix.c ../../../../../../rtems/c/src/../../cpukit/score/src/objectextendinformation.c: In function '_Objects_Extend_information': ../../../../../../rtems/c/src/../../cpukit/score/src/objectextendinformation.c:278:1: internal compiler error: Segmentation fault 0x93b6ef crash_signal ../../gcc/gcc/toplev.c:372 0x8e4890 reg_overlap_mentioned_p(rtx_def const*, rtx_def const*) ../../gcc/gcc/rtlanal.c:1519 0xc0dd5c xscale_sched_adjust_cost ../../gcc/gcc/config/arm/arm.c:11534 0xc0d97b arm_adjust_cost ../../gcc/gcc/config/arm/arm.c:11984 0xe1979f dep_cost_1(_dep*, unsigned int) ../../gcc/gcc/haifa-sched.c:1513 0xe1af07 dep_cost(_dep*) ../../gcc/gcc/haifa-sched.c:1532 0xe1af07 priority ../../gcc/gcc/haifa-sched.c:1683 0xe1afff set_priorities(rtx_insn*, rtx_insn*) ../../gcc/gcc/haifa-sched.c:6796 0x8ecd2c compute_priorities() ../../gcc/gcc/sched-rgn.c:2940 0x8ef94d schedule_region ../../gcc/gcc/sched-rgn.c:3033 0x8ef94d schedule_insns() ../../gcc/gcc/sched-rgn.c:3431 0x8f0053 schedule_insns() ../../gcc/gcc/sched-rgn.c:3416 0x8f0053 rest_of_handle_sched ../../gcc/gcc/sched-rgn.c:3635 0x8f0053 execute ../../gcc/gcc/sched-rgn.c:3743 Please submit a full bug report,
