This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Bootstrap failure on powerpc-ibm-aix4.1.4.0 after ppc branch merge
- To: geoffk at cygnus dot com
- Subject: Re: Bootstrap failure on powerpc-ibm-aix4.1.4.0 after ppc branch merge
- From: "Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>
- Date: Sun, 19 Mar 2000 09:33:45 -0500 (EST)
- Cc: egcs-bugs at egcs dot cygnus dot com, egcs-patches at egcs dot cygnus dot com
> From: Geoff Keating <geoffk@cygnus.com>
>
> > Date: Thu, 16 Mar 2000 17:31:01 -0500 (EST)
> > From: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu>
> > Cc: egcs-bugs@egcs.cygnus.com, egcs-patches@egcs.cygnus.com
> >
> > Now that the newppc branch is merged back in, I tried
> > bootstrapping on powerpc-ibm-aix4.1.4.0. The first problem I had was
> > that collect2 couldn't compile in stage1 because of missing functions
> > normally found in libld.a. I think x-aix41/x-aix43 used to have
> > CLIB=-lld in it, but these lines got moved to target files so cross
> > compiles would work. Now t-aix43 has the CLIB setting and there is no
> > t-aix41. So I created one. Here's the patch I used to get past this
> > problem.
> >
> > diff -rup orig/egcs-CVS20000316/gcc/config/rs6000/t-aix41 egcs-CVS20000316/gcc/config/rs6000/t-aix41
> > --- orig/egcs-CVS20000316/gcc/config/rs6000/t-aix41 Thu Mar 16 11:36:29 2000
> > +++ egcs-CVS20000316/gcc/config/rs6000/t-aix41 Thu Mar 16 10:55:55 2000
> > @@ -0,0 +1,4 @@
> > +# We need -lld for collect2 (actually this only matters
> > +# for a native compiler, but this is as good a place as any
> > +# to define the symbol).
> > +CLIB=-lld
> > diff -rup orig/egcs-CVS20000316/gcc/configure.in egcs-CVS20000316/gcc/configure.in
> > --- orig/egcs-CVS20000316/gcc/configure.in Thu Mar 16 08:33:34 2000
> > +++ egcs-CVS20000316/gcc/configure.in Thu Mar 16 10:57:28 2000
> > @@ -3157,7 +3157,7 @@ changequote([,])dnl
> > then
> > tmake_file=rs6000/t-xnewas
> > else
> > - tmake_file=rs6000/t-newas
> > + tmake_file="rs6000/t-newas rs6000/t-aix41"
> > fi
> > xmake_file=rs6000/x-aix41
> > float_format=none
>
> This is fine (with a suitable ChangeLog).
Thanks, installed (with, of course, a ChangeLog entry.)
> > Now having done that, I got an error in stage2 when building libgcc2.a:
> >
> > > _muldi3
> > > xgcc: Internal compiler error: program cc1 got fatal signal 4
> > > make[4]: *** [libgcc2.a] Error 1
>
> This is a generic bootstrap problem on ppc. I've been waiting until
> GCC is more stable before submitting the patch:
>
> Index: flow.c
> ===================================================================
> RCS file: /cvs/gcc/egcs/gcc/flow.c,v
> retrieving revision 1.235
> diff -p -u -c -r1.235 flow.c
> cvs server: conflicting specifications of output style
> *** flow.c 2000/03/15 20:24:35 1.235
> --- flow.c 2000/03/16 23:01:27
> *************** propagate_block (bb, old, significant, f
> *** 3302,3307 ****
> --- 3302,3315 ----
> PUT_CODE (next, NOTE);
> NOTE_LINE_NUMBER (next) = NOTE_INSN_DELETED;
> NOTE_SOURCE_FILE (next) = 0;
> +
> + if ((next = next_nonnote_insn (label)) != NULL
> + && GET_CODE (next) == BARRIER)
> + {
> + PUT_CODE (next, NOTE);
> + NOTE_LINE_NUMBER (next) = NOTE_INSN_DELETED;
> + NOTE_SOURCE_FILE (next) = 0;
> + }
> }
> }
> }
>
> --
> - Geoffrey Keating <geoffk@cygnus.com>
Thanks! It does indeed get me into bootstrap land.
--
Kaveh R. Ghazi Engagement Manager / Project Services
ghazi@caip.rutgers.edu Qwest Internet Solutions