[PATCH][ARM] bootstrap/38523 arm build fails to link cc1-dummy / configure tweak for stage1

Jakub Jelinek jakub@redhat.com
Sun Jan 11 17:23:00 GMT 2009


On Sun, Jan 11, 2009 at 05:51:35PM +0100, Laurent GUERBY wrote:
> On Sun, 2009-01-11 at 17:43 +0100, Paolo Bonzini wrote:
> > >> +elif test "x$enable_checking" = x; then
> > >> +  case "${host}" in
> > >> +    # Needed to avoid stage1 link failure, see PR bootstrap/38523 for details.
> > >> +    arm*-*-linux-gnueabi) stage1_checking=--enable-checking=release ;;
> > >> +    *) stage1_checking=--enable-checking=yes,types ;;
> > >> +  esac
> > >
> > > On behalf of people who patch the build system, I suggest an actual
> > > description of the problem.
> > 
> > I am not extremely fond of Laurent's patch.  
> 
> I have no preference between my two patches :).
> 
> > I'd have preferred -O1,
> > as --enable-checking=release may indeed hide some bugs.  I didn't
> > complain because ARM crosses are built more often than ARM native
> > compilers, because an ARM host does not have anything particularly
> > peculiar (IIRC it has unsigned chars, but powerpc-linux also has
> > them), and because 25% slowdown is a problem.  (But then, there has
> > been talk for years of enabling -O1 for stage1, on the ground that it
> > might improve build times... and now 25% slowdown?  Heck...)
> 
> The slowdown is -O1 vs -O0 --enable-checking=release (since -O0 alone
> doesn't bootstrap) so it's not a real indication of -O1 vs -O0.
> 
> If there's a preference on -O1 (and debian builds use -O1) then let's
> go with it, the arm native machines are slow so no one expects
> fast builds anyway.

Or force -O1 in config.host only if:
case "${host}" in
arm-*-linux-*eabi)
  case "$ac_checking_flags" in
    no)
    release)
      ;;
    *)
      host_xmake_file="$host_xmake_file x-cflags-O1"
      ;;
  esac
  ;;
esac
Then people can choose, either they build with --enable-stage1-checking=release
(or build release instead of devel build) and have faster, but less
checking, bootstrap, or they do the default, slightly slower, but safer.

	Jakub



More information about the Gcc-patches mailing list