http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53555
--- Comment #4 from H.J. Lu <hjl.tools at gmail dot com> 2012-06-02 12:07:59 UTC --- (In reply to comment #3) > Hmm, it seems it's not quite correct to check ix86_tune, proper way is to use > ix86_arch here. Following change fixes he bootstrap > > diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c > index c6c2c5e..a272411 100644 > --- a/gcc/config/i386/i386.c > +++ b/gcc/config/i386/i386.c > @@ -23881,7 +23881,7 @@ ix86_sched_reorder(FILE *dump, int sched_verbose, rtx > *ready, int *pn_ready, > issue_rate = ix86_issue_rate(); > > /* Do reodering for Atom only. */ > - if (ix86_tune != PROCESSOR_ATOM) > + if (ix86_arch != PROCESSOR_ATOM) > return issue_rate; > /* Nothing to do if ready list contains only 1 instruction. */ > if (n_ready <= 1) This is the wrong fix. Bootstrap will fail if GCC is configured with --with-arch=atom.