From mboxrd@z Thu Jan  1 00:00:00 1970
From: Johannes Stezenbach <js@sig21.net>
Date: Wed, 30 Jan 2008 16:36:04 +0100
Subject: [U-Boot-Users] [Patch] Disable icache before call the
	first	line of kernel in do_bootelf().
In-Reply-To: <200801291912.11046.vapier@gentoo.org>
References: <20080129223551.3A9EA24680@gemini.denx.de>
	<200801291912.11046.vapier@gentoo.org>
Message-ID: <20080130153604.GA9497@sig21.net>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

On Tue, Jan 29, 2008 at 07:12:10PM -0500, Mike Frysinger wrote:
> On Tuesday 29 January 2008, Wolfgang Denk wrote:
> > In message <200801282114.14393.vapier@gentoo.org> you wrote:
> > > unfortunately, using weak symbols and overriding elsewhere doesnt look
> > > like it's possible currently due to the way ld searches archives.  for
> > > example, if
> >
> > ???
> >
> > > ld will pick the weak symbol provided by libcommon.a even though a strong
> > > symbol is also available in libblackfin.a :(
> >
> > That should never happen. What is your toolchain?
> 
> read the binutils mailing list.  this is the expected behavior of ld.
> 
> http://sourceware.org/ml/binutils/2008-01/msg00301.html

Overriding weak with strong symbols works just fine, however
in your case ld has no reason to even look at your override,
since it already has a (weak) definition for do_bootelf_setup().

Thus the solution for your problem is to give ld a reason
to pull in your do_bootelf_setup() definition. There are
two ways:

- explicitly list the .o file on the linker command line,
  i.e. add it to $(OBJS) or $(PLATFORM_LIBS) instead
  of libblackfin.a

- put the do_bootelf_setup() definition in a .o file
  along with some other code you know will be pulled in,
  e.g. add it to lib_blackfin/cache.c instead of putting
  it in its own file


HTH,
Johannes