Power Host Machine

J_Man jeremy at jutley.org
Thu Jan 23 01:16:16 PST 2003


Brenton Schulz wrote:
> Rob Millerd wrote:
> 
>> Just wondering if it's OK to have a powerful host machine and then move
>> the hard drive to an older computer.  For example I have a PIII 730MHz
>> with 256 megs of ram as my host computer.  I'm building a LFS system to
>> be a dial-up router computer to run on an old 386 133MHz with 16megs of
>> ram.  I noticed while installing Binutils that it said i686.  does this
>> matter?  Will I have any problems running LFS on the old computer in
>> this situation?
>>
>> Thanks  
>>
> More than likely yes. The i686 extensions will be illegal instructions 
> to the 386 CPU and hence the binary files will not run. Once you've done 
> the configure's on each of the programs to be compiled, edit the 
> makefile's and look for where it's telling gcc to compile for i686 and 
> change it to i386.
> 
> someone else can fill in the gritty details :p
> 

There's an even easier solution to the i686 compiling for a 386 dilema. 
  Before you chroot, do this:

cp $LFS/static/bin/uname $LFS/static/bin/uname.orig

cat > $LFS/static/bin/uname << "EOF"
#!/static/bin/bash
/static/bin/uname.orig "$@" | sed "s/i686/i386/"
EOF

chmod 755 $LFS/static/bin/uname

Then, when you enter chroot, everything will THINK you have a 
super-powered 386 box.  Begin chap 6 as indicated by the book.  After 
you compile sh-utils, you'll need to make the same hack to the uname it 
installs:

cp /bin/uname /bin/uname.orig


cat > /bin/uname << "EOF"
#!/static/bin/bash
/bin/uname.orig "$@" | sed "s/i686/i386/"
EOF

chmod 755 /bin/uname


Read the cross-compiling-x86 hint for more details.

Jeremy

-- 
I've heard it said that there are two types of *NIX admins, those that
have done something EXTREMELY stupid while logged in as root, and
those that WILL do something extremely stupid while logged in as root.

-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-support' in the subject header of the message



More information about the lfs-support mailing list