i just want to clarify the apparent relationship between header
files. header files are, for the most part, either arch-independent
(include/linux/*.h), or arch-specific (include/asm-<whatever>/*.h).
as i see it, a header "file" can fall into one of three categories:
1) purely arch-independent (say, include/linux/kobject.h)
2) purely arch-specific (include/asm-*/semaphore.h)
3) a combination of 1) and 2), where there is an independent
version of the file, which contains some independent
content and, furthermore, includes the arch-specific version
of the header file, say:
include/linux/uaccess.h
include/asm-*/uaccess.h
is that about right? i ask since that suggests the obvious way to
move some header info around to make it more organized.
for instance, include/linux/kernel.h contains macro definitions for
BUILD_BUG_ON() and BUILD_BUG_ON_ZERO(). however, there are a number
of arch-specific "bug.h" files that, for each arch, define the macro
BUG().
would it not make more sense to move those arch-independent macro
defines out of kernel.h and into include/linux/bug.h? that way, all
the "bug"-related stuff is now available through the single include
#include <linux/bug.h>
and, obviously, that idea holds for more than just that example.
thoughts?
rday
--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive: http://mail.nl.linux.org/kernelnewbies/
FAQ: http://kernelnewbies.org/faq/
