[Clfs-dev] powerpc64 and sparc64 discussion for 32 bit user space and multilib

[William Harrington]

On Dec 17, 2008, at 7:32 PM, William Harrington wrote:

> Okay, I'm now up to the final system build. My question(s) follow:
>
> For just 32 bit user space is it wise to build (along with glibc)  
> multilib gmp and mpfr prior to the gcc build?
>
> My plan was to go strictly 32 bit using the uname_hack where uname - 
> m outputs ppc after I built gcc.
>
> More input is needed. I think it is wise to build multilib up to gcc  
> then load the uname_hack and then
> build using the ppc book after gcc is build and installed.
>
> -William

Another thought....... or shall there be no 64 bit at all, and just 32  
bit user space after chrooting.

If that is the case, then load the uname_hack at the start of chap10.

mpfr tests will fail, but I think that is normal when testing 32bit  
with a 64 bit machine.

My understanding was that there would be 32 bit and 64 bit toolchain  
available, but nothing after gcc would be 64 bit, just 32 bit.  
Although, I'm entering some issues that might be my fault or not. I'm  
gonna redo the process tomorrow and fix my notes. Although, if the  
idea is to have nothing 64 bit in the system, just the gcc to build a  
64 bit kernel, then the whole problem is solved, as far as I know,  
with my build process.

gcc now defaults to 32 bit at least! The other issue is when using  
this on a ppc64 machine, that config.guess will return powerpc64- 
unknown-linux-gnu, which will cause problems, that's why the  
uname_hack has entered the world of my build.

-William