This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: gold bug combining .bss sections...
From: Ian Lance Taylor <iant@google.com>
Date: Fri, 11 Apr 2008 17:33:38 -0700
> David Miller <davem@davemloft.net> writes:
>
> >> David Miller <davem@davemloft.net> writes:
> >>
> >> > If, for example, we have a normal .bss in the final link with only a
> >> > necessary alignment of 4, and we also have a dynbss .bss with a copy
> >> > reloc needing an alignment of 8, gold outputs the final combined .bss
> >> > section with only an alignment of 4.
> >>
> >> I just committed this patch which I think should fix this problem.
> >
> > Unfortunately it does not. (I remembered to remove the
> > .bss alignment hack in my sparc copy reloc function before
> > testing).
>
> Hmmmm, well, it does seem to fix it in the limited sense that the
> alignment of the .bss section in two_file_shared_2_test is 8, whereas
> you indicated that it was previously 4. That is, I suspect that now
> something else is failing.
>
> Since I don't have easy access to a SPARC system, how does the new
> binary fail?
t17data is not 8 byte aligned, so we take a SIGBUS.
t17data is a copy reloc which needs to be 8 byte aligned
but it isn't.
Perhaps when the various .bss section data are combined, things that
were aligned end up not being so.