Re: binutils-2.16.1

[email protected] (Arthur Othieno)
Newsgroups gmane.linux.ports.game-cube.devel
Message-ID <20050716020812.GH2399@mars>
On Sat, Jul 16, 2005 at 01:29:38AM +0200, Groepaz wrote:
> On Saturday 16 July 2005 00:57, Arthur Othieno wrote:
> 
> > And so is buildroot[1]. The uClibc toolchain is simply pre-configured
> > to target ppc. Apart from that, nothing fancy.
> >
> > AFAIK, no one has slapped the "official" moniker over it. Allow me to
> > quote myself[2]:
> >
> >   "That said, the recommended compiler, for both kernel and userspace,
> >    is the uClibc one: http://uclibc.org/toolchains.html"
> 
> ...which is probably one of the reasons of little interest in gc-linux among 
> the ppl in the gcdev channel. i've personally dropped it myself coz when i 
> complained a while back (even on this list) that gc-linux cant be compiled 
> with a "clean" gcc, ppl would keep telling me to use that uclibc toolchain 
> instead of fixing whatever is borked in the makefiles. (it were trivial 
> things, that much i remember)
 
Nothing is broken in the Makefiles. The only thing that was "broken" was the
top-level Makefile futzing directly with $(CROSS_COMPILE) and thus making it
toolchain-biased, so to speak. This has since been deleted.

ISTR your problem was __powerpc__ not being defined. I pointed out that
this was due to a misconfigured compiler, and to which you acknowledged[1].

In any case, if your version of gcc fails this test, then reconfiguring
gcc with --target=powerpc-foo-linux should do the trick[2].

  hubb@mars:~$ powerpc-linux-gcc -dumpversion
  3.3.3
  hubb@mars:~$ echo foo | powerpc-linux-gcc -dM -E - | grep powerpc
  #define __powerpc__ 1
  #define __powerpc 1
  #define powerpc 1
  hubb@mars:~$

> > > i personally never understood why the gc-linux ppl made their own stuff
> > > in that area. maybe some ppl should visit the gcdev irc channel on efnet
> > > more often :=P
> >
> > Ok, this sounds a little too religious. Besides, we are talking about
> > _binutils_ here, and not about (a collective) toolchain $foo vs.
> > toolchain $bar. That alone is another debate of it's own, and one
> > in which I have no interest participating.
> 
> imho its not very smart to use seperate distributions of binutils und gcc, 
> update them seperatly and all that, its just asking for trouble. and there 
> are reasons for so many ppl to use devkitpro - it just works damn well.
> 
> > I use what works for me, and I'm pretty sure everyone else uses what
> > works for _them_. It's only natural.
> 
> ofcourse, no problem with that. the point when its starting to get bad is when 
> someone relies on certain specific things that are only present in a specific 
> setup - like the gclinux makefiles do.
 
Again, none of the Makefiles in the kernel tree are toolchain-specific.
Feel free to point out any specific cases.


[1] http://article.gmane.org/gmane.linux.ports.game-cube.devel/235
[2] http://ozlabs.org/pipermail/linuxppc-embedded/2004-February/013183.html


-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.