A query.. Build LFS for a different target?

Ronald Hummelink maillist at hummelink.xs4all.nl
Sat Nov 8 08:36:27 PST 2003

On Sat, 2003-11-08 at 17:23, Bill's LFS Login wrote:
> On Sat, 8 Nov 2003, Dale Weber wrote:
> > On Saturday 08 November 2003 2:05 am, Ronald Hummelink wrote:
> >
> > > thats a whole different thing. I do not believe there is a hint about
> > > it. There is a hint about building a toolchain targetting cygwin/mingw32
> > > but that is not likely the same as eg sparc64.
> >
> > 	Actually, I didn't mention a specific non-x86 architecture, but am thinking
> > more along the lines of smaller  embedded processors like the Intel XScale.
> >
> > 	It looks like more research into this is needed, but -seems- like it -should-
> > be possible to build an LFS from a host system (like x86) for a completely
> > different target..  I think the way the LFS is built is basically the same
> > way embedded Linux is built for these processors, but I could be way off base
> > too.
> I may be off-base here, but I seem to recall some words in the "info
> gcc" that address the issues of concern. The compiler has a concept of 3
> (IIRC) compiler parameters: host, target and build(?). I don't remeber
> any details, but it sounded like that was the way to make a compiler on
> one machine produce code for another, regardless of architecture.
> Info gcc may lead you down the road (to confusion?).  :)

It is more complicated then that;
The LFS build method assumes you can RUN the build binaries on the
builders (host in gcc terms) system.
If you want to compile a system for $target then you normally cannot run
the $target binaries so you need to use a different method using a cross
toolchain.. which is outside the scope of (current) LFS imho :)


> >
> > 	8-Dale
> >
> -- 
> Bill Maltby
> lfsbillATwlmcsDOTcom
> Fix line above & use it to mail me direct.
Linux is like a small snowball rolling downhill. Microsoft is just
waiting down the mountain...

More information about the lfs-support mailing list