svn commit: r201815 - head

[M. Warner Losh]

In message: <20100110.001429.173527971.nyan at FreeBSD.org>
            TAKAHASHI Yoshihiro <nyan at freebsd.org> writes:
: In article <201001081901.o08J1pjc062909 at svn.freebsd.org>
: "Bjoern A. Zeeb" <bz at FreeBSD.org> writes:
: 
: >   To avoid hardcoding further kernel configuration names for
: >   make universe, split the logic into two parts:
: >   - 1st to build worlds and generate kernel configs like LINT.
: >   - 2nd to build kernels for a given TARGET architecture correctly
: >     finding all newly generated configs, not knowing anything about
: >     LINT anymore. (*)
: 
: > +universe_kernels: universe_kernconfs
: > +BUILD_ARCH!=	uname -p
: 
: Should this be "uname -m"?
: 
: 
: > +TARGET?=	${BUILD_ARCH}
: > +KERNCONFS!=	cd ${.CURDIR}/sys/${TARGET}/conf && \
: > +		find [A-Z0-9]*[A-Z0-9] -type f -maxdepth 0 \
: > +		! -name DEFAULTS ! -name NOTES

Yes.  Also, the name BUILD_ARCH is misleading, since it corresponds to
MACHINE, not to MACHINE_ARCH.  it is clear that it is used in the
MACHINE context, not the MACHINE_ARCH context.

These two are often confused, but many of the embedded platforms have
added to the confusion and it is unclear where the sharp dividing line
should be here anymore.  It used to be separate boot environment, but
even that has gotten fuzzy for architectures like mips and arm that
boot with a huge variety of boot loaders...

Warner