This is the mail archive of the newlib@sources.redhat.com mailing list for the newlib project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: powerpc-eabism, strategy for finding crt0.o


> Cc: Geoff Keating <geoffk@redhat.com>, bkoz@redhat.com,
>         gcc-patches@gcc.gnu.org, newlib@sources.redhat.com
> From: Alexandre Oliva <aoliva@redhat.com>
> Organization: GCC Team, Red Hat
> Date: 13 Apr 2001 17:34:02 -0300
> User-Agent: Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.1 (Cuyahoga Valley)
> 
> On Apr 13, 2001, Michael Meissner <meissner@cygnus.com> wrote:
> 
> > The current powerpc behavior is if you don't specify -msim/-mvme/etc. and
> > configured it for powerpc-eabi (not powerpc-eabisim), it will link in no
> > library (ie, no -lc, no -lvme, etc.) or crt0.
> 
> Hmm...  This is bad.  Are there more ports like this?  I don't see a
> way to get autoconf link tests to work on such platforms.

Yup.  I believe that MIPS is like this.

Yes, you cannot use autoconf link tests on such platforms.  You have
to deal with it in some other way.

If you really must know whether some link test will pass, you have to
break the configuration up into smaller pieces, eg.

powerpc-eabisim
powerpc-vxworks
powerpc-eabiads
powerpc-eabiyellowknife
powerpc-linux
powerpc-eabimvme

but even then, I don't know how you can do link tests on (for
instance) vxworks, where final link takes place on the target board.
And then, once you've done this, you have a toolchain that will only
work for the one board you've targetted it at.

MIPS is even worse, in that I don't think that you can as easily build
compilers that default to particular boards.  On ppc, it's just a
question of supplying a config file that sets the default, but on MIPS
you have to actually write the specs.

-- 
- Geoffrey Keating <geoffk@geoffk.org>


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]