This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Fix libjava x86-64 multilib
Alexandre Oliva <aoliva@redhat.com> writes:
> On Oct 1, 2002, Bo Thorsen <bo@sonofthor.dk> wrote:
>
>> This fixes the multilib error I've been getting. Bootstrapped on x86-64
>> (finally!). Can I commit it?
>
> Wow! Is LD really used by libtool on x86_64*linux*? I'm surprised.
On all platforms, LD is used for relocatable links (ld -r) :-(
> Do you have a pointer to the message in which you give more details
> about the problem? I must have missed it :-(
It went over the GCC list:
http://gcc.gnu.org/ml/gcc/2002-10/msg00009.html
http://gcc.gnu.org/ml/gcc/2002-10/msg00022.html
http://gcc.gnu.org/ml/gcc/2002-10/msg00039.html
>
>> 2002-10-01 Bo Thorsen <bo@suse.de>
>
>> * configure: rebuilt after toplevel libtool.m4 fix.
>
>> 2002-10-01 Bo Thorsen <bo@suse.de>
>
>> * libtool.m4: Fix x86-64 multilib
>
> Please make sure this makes it to libtool CVS if it's not there yet.
> As soon as it makes it there, it's ok to install it in GCC and src
> too.
>
>> - alpha* | hppa* | i*86 | powerpc* | sparc* | ia64* )
>> + alpha* | hppa* | i*86 | powerpc* | sparc* | ia64* | x86_64 )
>> lt_cv_deplibs_check_method=pass_all ;;
>
> Are you 100% sure about this, i.e., that all non-PIC x86_64
> relocations can be turned into dynamic relocations that the dynamic
> linker can resolve. I seem to recall having heard something contrary
> to this, but I'm no expert on x86_64. If this is really safe for
> x86_64, this chunk is ok too, after it makes it to libtool CVS, but
> the ChangeLog should mention this change.
I don't fully understand your answer but we need -fPIC for shared libs.
Andreas
>
>
> If you have trouble getting the patch approved for libtool CVS, please
> get in touch with me in private and I can try to help you out.
>
>
> Thanks,
--
Andreas Jaeger
SuSE Labs aj@suse.de
private aj@arthur.inka.de
http://www.suse.de/~aj