This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: PPC relocs in shared libraries
- From: Daniel Jacobowitz <drow at mvista dot com>
- To: Richard Henderson <rth at redhat dot com>, Ian Lance Taylor <ian at zembu dot com>,Jakub Jelinek <jakub at redhat dot com>, Geoff Keating <geoffk at geoffk dot org>,binutils at sources dot redhat dot com
- Date: Tue, 1 Oct 2002 09:43:26 -0400
- Subject: Re: PPC relocs in shared libraries
- References: <20020925220209.GA8201@nevyn.them.org> <200209252309.g8PN9YG18648@desire.geoffk.org> <20020926193900.GA31145@nevyn.them.org> <20020927144437.H14457@bubble.sa.bigpond.net.au> <20020927130435.GA3791@nevyn.them.org> <20020927090755.E21220@devserv.devel.redhat.com> <20020927165404.GH19327@redhat.com> <20021001171344.N25369@bubble.sa.bigpond.net.au>
On Tue, Oct 01, 2002 at 05:13:44PM +0930, Alan Modra wrote:
> On Fri, Sep 27, 2002 at 09:54:04AM -0700, Richard Henderson wrote:
> > On Fri, Sep 27, 2002 at 09:07:55AM -0400, Jakub Jelinek wrote:
> > > I still don't understand why GDB needs to care about reloc sections in
> > > ET_DYN/ET_EXEC.
> >
> > The relocation is the only way to associate the real address with the
> > debug info in the case of dynamic linker overrides.
>
> Hmm, OK. ld.so won't do the reloc for gdb though, will it? So gdb
> will need to be quite clever about symbol resolution. No reason not
> to emit the relocs though, except to reduce file size.
Which it isn't. If you have multiple functions with the same name GDB
is prone to getting helplessly confused. Anyone who's had to debug,
say, elfcode.h in a multi-target binutils is well aware of this :)
--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer