This is the mail archive of the
mailing list for the glibc project.
Re: libgcc-compat for linux-mips
On Sun, Oct 20, 2002 at 02:55:48PM -0700, H. J. Lu wrote:
> On Sun, Oct 20, 2002 at 09:41:12PM +0000, Guido Guenther wrote:
> > On Sun, Oct 20, 2002 at 02:26:23PM -0700, H. J. Lu wrote:
> > > On Sun, Oct 20, 2002 at 11:18:28PM +0200, Guido Guenther wrote:
> > > > On Sun, Oct 20, 2002 at 02:08:24PM -0700, H. J. Lu wrote:
> > > > > > So why am I not seeing __udivdi3 as undefined symbol here? Any ideas?
> > > > >
> > > > > __udivdi3 is defined in /lib/libuuid.so.1:
> > > > > binding file /lib/libuuid.so.1 to /lib/libuuid.so.1: normal symbol `__udivdi3'
> > > > No - not after recompiling in a gcc-3.2 & glibc 2.3.1 chroot:
> > > > nm -D --defined-only /lib/libuuid.so.1 | grep __u | wc -l
> > > > 0
> > >
> > > I haven't qualified gcc 3.2 for Linux/mips yet. It is on my todo list
> > > as the part of my RedHat 8.0/mips port.
> > I still don't understand how these symbols get resolved. In case we
> > should need a libgcc-compat, I have a patch prepared at:
> > http://honk.physik.uni-konstanz.de/linux-mips/glibc/patches/experimental/libgcc-compat-mips.dpatch
> > It contains all the symbols found scanning the whole debian archive
> > (thanks to Jack's script).
> Please make sure your binutils is up to date. I got
> # readelf -a mke2fs| grep __udivdi3
> 114: 00409ac0 0 FUNC GLOBAL HIDDEN 11 __udivdi3
> 141: 00409ac0 0 FUNC GLOBAL HIDDEN 11 __udivdi3
> The current one is the Linux binutils 220.127.116.11.11.
Ooops. I meant to say 18.104.22.168.10.