http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48442
--- Comment #2 from Alexander Monakov <amonakov at gcc dot gnu.org> 2011-04-06 13:41:33 UTC --- Confirmed. The 4.4/4.5 failure is most likely a dup of PR 46204. The 4.6/4.7 failure is due to selective scheduler removing a conditional jump during purge_empty_blocks in sel_sched_region_1 and then tripping on an incorrect assert in init_seqno (we probably never managed to remove jumps at that point prior to PR 46204 fix, but the nearby comment suggests we could and contradicts the assert). I'll test the attached patch that relaxes the assert and also drops number_of_insns argument to clean up a bit.
