------- Additional Comments From uros at kss-loka dot si 2005-04-14 05:55 ------- I susupect the bug to be in expand_builtin_int_roundingfn(). It looks that
tmp = gen_reg_rtx(floatmode); is needed before tmp = expand_builtin_mathfn (exp, NULL_RTX, NULL_RTX); It is somehow annoying fact that everything works OK on i686-pc-linux-gnu. -- What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|unassigned at gcc dot gnu |uros at kss-loka dot si |dot org | Status|NEW |ASSIGNED Last reconfirmed|2005-04-14 01:27:58 |2005-04-14 05:55:21 date| | http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21004