This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: V3 vs. AIX
- To: dje at watson dot ibm dot com
- Subject: Re: V3 vs. AIX
- From: Mark Mitchell <mark at codesourcery dot com>
- Date: Mon, 05 Feb 2001 10:31:52 -0800
- Cc: bkoz at nabi dot net, gcc at gcc dot gnu dot org
- Organization: CodeSourcery, LLC
- References: <mark@codesourcery.com><200102051813.NAA27176@mal-ach.watson.ibm.com>
>>>>> "David" == David Edelsohn <dje@watson.ibm.com> writes:
David> Which GAS are you using? The one in the gdb-5 CVS sources or the
David> one in the AIX "Linux Affinity" release?
From CVS.
David> My suggestion always has been to use GAS for just ios.cc and
David> string-inst.cc, not everything (:^).
True. I've been hoping to avoid making things any more complicated
than they are, but your procedure is worth a try.
David> Neither of those addresses in r2 or r12 look particularly
David> correct. I am guessing that TOC slot 608 loaded into r12
did not receive
David> a fixup and the value is the unchanged offset (AIX relocations are
David> relative offsets, not absolute substitutions, so that value
probably is
David> the displacement from the beginning of the section which
really contains the symbol).
That make sense -- but it doesn't seem to correspond to what's
actually going on. At least I can't make those numbers work out.
David> Because of System V braindamage we are trying to emulate, we are
David> linking with the -berok flag, so I am a little worried that this is a
David> symbol we allowed the linker to leave unresolved.
I'm slightly confused by the `.' in front of the name, but the symbol
(with the `.') appears *both* as a defined function:
000000010069300 <._ZNSs10_M_replaceIPKcEERSsSt17__normal_iteratorIPcSsES5_T_S6\_St20forward_iterator_tag>:
10069300: 7c 08 02 a6 mflr r0
10069304: 93 e1 ff fc st r31,-4(r1)
and as a TOC glue thingy:
000000001004b480 <._ZNSs10_M_replaceIPKcEERSsSt17__normal_iteratorIPcSsES5_T_S6\_St20forward_iterator_tag>:
1004b480: 81 82 02 5c l r12,604(r2)
1004b482: R_TOC _ZNSs10_M_replaceIPKcEERSsSt17__normal_\iteratorIPcSsES5_T_S6_St20forward_iterator_tag+0xffffffffdffe16ec
Should one of these not have the `.'? Would that be consistent with
your undefined symbol hypothesis?
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com