On Thu, Aug 2, 2012 at 1:57 AM, Julian Brown <jul...@codesourcery.com> wrote: > FWIW, I've hit this issue in the past, and used a patch as follows to > fix it: > > @@ -12015,7 +12025,10 @@ create_fix_barrier (Mfix *fix, HOST_WIDE > gcc_assert (GET_CODE (from) != BARRIER); > > /* Count the length of this insn. */ > - count += get_attr_length (from); > + if (LABEL_P (from) && (align_jumps > 0 || align_loops > 0)) > + count += MAX (align_jumps, align_loops); > + else > + count += get_attr_length (from); >
That looks like it's from a version of arm.c before get_label_padding was added. The existing trunk code uses get_label_padding at that spot, and my change was to get_label_padding. > Not sure if that's equivalent, but you might want to check > [-falign-loops] too while you're at it. That seems like it could overestimate, if the -falign-loops value is larger than the -falign-labels value. Not every label is for a loop. Thanks, Roland
