This is the mail archive of the
glibc-linux@ricardo.ecn.wfu.edu
mailing list for the glibc project.
Re: Difficulties with cross compiling glibc for ARM.
- To: glibc-linux at ricardo dot ecn dot wfu dot edu
- Subject: Re: Difficulties with cross compiling glibc for ARM.
- From: Steve Hill <sjhill at cotw dot com>
- Date: Fri, 09 Jun 2000 08:02:27 -0500
- References: <393FC59B.11DAB609@cotw.com> <20000609003552.A15361@baoule.ath.cx>
- Reply-To: glibc-linux at ricardo dot ecn dot wfu dot edu
Gael Queri wrote:
>
> > /opt/build-glibc/libc.so.6: undefined reference to `__ashldi3'
> > /opt/build-glibc/libc.so.6: undefined reference to `__divdi3'
> > /opt/build-glibc/libc.so.6: undefined reference to `__umoddi3'
> > /opt/build-glibc/libc.so.6: undefined reference to `__udivdi3'
> > /opt/build-glibc/libc.so.6: undefined reference to `__lshrdi3'
> > /opt/build-glibc/libc.so.6: undefined reference to `__moddi3'
> > /opt/build-glibc/libc.so.6: undefined reference to `__muldi3'
>
> Well, these are auxiliary functions of gcc to play with long long which
> must be in libgcc.a and it seems that your libgcc.a doesn't have these.
> How did the gcc compile went? Can you try to rebuild libgcc and see if
> something changes? (use nm to see if these are present)
>
Thanks for the reply. Since I am building a cross compilation toolchain,
the above symols are not placed in my 'libgcc.a' when only building the
cross GCC for the first time. They don't appear until after I rebuild gcc
when I have a working glibc. Thus you see the problem. This is frustrating
considering I did this once before (a long time ago) and things worked just
fine. I did a 'arm-linux-nm' on my 'libgcc.a' and only the `__XXXXsi3'
symbols are present. Any further thoughts.
-Steve
--
Steven J. Hill - Embedded SW Engineer
Public Key: 'finger sjhill@mail.cotw.com'
FPR1: E124 6E1C AF8E 7802 A815
FPR2: 7D72 829C 3386 4C4A E17D