[Clfs-support] Concerned about triplet in 'chroot'

Dan McGhee beesnees at grm.net
Sat Sep 5 15:13:19 PDT 2009

Using SVN-20090902-Native-Multilib

Finshed the temporary system and mounted the virual kernel filesystems.  
Came to "All Stop" in Ch. 6.33 "Before Entering the Chroot 
Environment."  Obviously the moral of the story is "make sure the 
triplet is right or it'll screw everything up," but I see conflicts with 
my various outputs and the directions.  There's a thread in the archives 
that only served to muddy the waters for me--"[Clfs-support] 64 bit 
Source machine".  The help in that thread seemed to be "this is a 
problem only on i[3-6]86 machines.  I have x86_64.

<echo ${CLFS_TARGET}> =="x86_64-pc-linux-gnu"

from automake-1.11 <lib/config.guess=="x86_64-unknown-linux-gnu"

and, finally <uname -m>=="x86_64"

Book says, "If the output of that command does not equal what is in 
${CLFS_TARGET} then you need to read on."  Obviously the output of the 
command and the value of the variable ALMOST match and 'uname -m' gives 
64 bit, but not any of the rest.  But is this good enough to proceed?

When I "read on," I find the test using setarch and ran it.

<setarch linux32 lib/config.guess>=="i686-pc-linux-gnu"

Then the book says "If the output of the command above--the setarch 
linux32--equals what is in ${CLFS_TARGET} then you have a viable 
solution."  But the output of this command is completely different than 
${CLFS_TARGET}.  So I read further and encounter a kernel module with 
which I can "fool" the system into using "i686," which doesn't seem 
reasonable for 64 bit machine, running a 64 bit kernel from Ubuntu.

Unless I botched setting up CLFS_TARGET  I'm really at a loss on what to 
do next.  I would surely appreciate some guidance.  Maybe the "multilib" 
has me confused, but isn't that what BUILD{32,64} does?


More information about the Clfs-support mailing list