Re: bootstrap broken on mainline for targets not using DFA

[Vladimir Makarov <vmakarov at redhat dot com>]

Mostafa Hagog wrote:

> Vladimir Makarov <vmakarov@redhat.com> wrote on 08/06/2004 23:24:59:
>
>  
>
> > Bernardo Innocenti wrote:
> >
> >    
> >
> > > Mostafa Hagog wrote:
> > >
> > >      
> > >
> > > > ChangeLog
> > > > 2004-05-23  Ayal Zaks  <zaks@il.ibm.com>
> > > >      Mostafa Hagog  <mustafa@il.ibm.com>
> > > >
> > > >      * Makefile.in (modulo-sched.o, ddg.o): New.
> > > >      * ddg.h, ddg.c, modulo-sched.c: New files.
> > > >      [...]
> > > >        
> > >
> > > This patch breaks bootstap on the few remaining backends that
> > > don't yet use the DFA scheduler, notably m68k:
> > >      
> > It will happen for all ports not having a pipeline description (DFA or
> > old one).
> >    
>
> If so, shouldn't we use the TARGET_SCHED_USE_DFA_PIPELINE_INTERFACE macro
> instead ?
>
>  
There is no such macro anymore.  A hook is used for this.  I think the 
usage of INSN_SCHEDULING is perfectly fine.

The hook is already used in modulo-sched.c.  If it returns 0, the code 
of modulo scheduling will be not run.

Vlad