Re: -dlopen self and optimizing away symbols

[Albert Chin]

On Thu, Aug 17, 2006 at 07:42:00AM +0200, Ralf Wildenhues wrote:
> * Albert Chin wrote on Thu, Aug 17, 2006 at 12:39:25AM CEST:
> > Just built openldap-2.3.25 on HP-UX 11.23/IA. 'make check' fails
> > because a symbol wasn't available in the main binary which dlopen()'ed
> > a module that required the symbol. The main binary was built with
> > "-dlopen self".
> 
> > Unfortunately, the HP-UX 11.23/IA compiler, when generating
> > .libs/slapd, will optimize _away_ symbols that are not used by any of
> > the object files being linked:
> 
> Which compiler option causes it to optimize away the symbols?

+O2. +O1 and no optimization seem to be ok.

> Which linker option is that translated to?

+O2 passes +Oprocelim to ld.

> > How do we solve this? I suppose we could build without CFLAGS but then
> > we'd lose +DD64 for 64-bit objects.
> 
> Probably adding a linker flag at the end to undo the dangerous setting
> should be safest, given there is such a flag.
> 
> But say, why is -Wl,-E not passed to the link line?  It should be
> export_dynamic_flag_spec.  And it should have the desired effect:
> |  -E             Mark all symbols defined by a program for export
> |                 to shared libraries.  In a +compat mode link, ld
> |                 marks only those symbols that are actually
> |                 referenced by a shared library seen at link time.
> |                 In a +std link, all symbols are exported by
> |                 default, so -E is not necessary to make symbols
> |                 visible.  However, it has an additional side
> |                 effect of identifying all exported symbols as
> |                 necessary, so that they will not be removed when
> |                 using dead code elimination (+Oprocelim).

Because -export-dynamic is not specified.

> Otherwise, -Wl,+Onoprocelim should be the hammer we're looking for, I
> guess; but that should not even be necessary, given above documentation
> (from ld(1)).

Well, export_dynamic_flag_spec isn't even looked at when -dlopen self
is being handled. Maybe it should and only use the -dlopen self code
when export_dynamic_flag_spec is empty?

If not, looks like we'll need to add +Onoprocelim to the cc
command-line for HP-UX. What's the best way to do this? A new
variable?

-- 
albert chin (address@hidden)