https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116189
--- Comment #12 from Oleg Endo <olegendo at gcc dot gnu.org> --- (In reply to Andrew Pinski from comment #8) > ``` > (gdb) condition 1 x_rtl.emit.x_cur_insn_uid==1083 > > #0 make_insn_raw (pattern=0x7ffff7570858) at ../../gcc/emit-rtl.cc:4137 > #1 0x0000000001f10469 in sh_recog_treg_set_expr (op=0x7ffff7570018, > mode=E_SImode) at ../../gcc/config/sh/sh.cc:12300 > #2 0x000000000280042c in treg_set_expr (op=0x7ffff7570018, mode=E_SImode) > at ../../gcc/config/sh/predicates.md:628 > #3 0x0000000001ef5462 in addsubcosts (x=0x7ffff7570060) at > ../../gcc/config/sh/sh.cc:3177 > #4 0x0000000001ef5f7f in sh_rtx_costs (x=0x7ffff7570060, mode=E_SImode, > outer_code=25, opno=1, total=0x7fffffffd2a4, speed=false) at > ../../gcc/config/sh/sh.cc:3466 > #5 0x000000000197ead3 in rtx_cost (x=0x7ffff7570060, mode=E_SImode, > outer_code=SET, opno=1, speed=false) at ../../gcc/rtlanal.cc:4605 > ``` > > oh this is a bad hack but ok. Please forgive me, I was probably desperate at that time. Thanks a lot for figuring this out!
