should include/uapi include header files *unused* in kernel space?

[Robert P. J. Day]

On Fri, 3 Feb 2017, Jim Davis wrote:

> On Sun, Jan 22, 2017 at 6:18 AM, Robert P. J. Day <rpjday at crashcourse.ca> wrote:
> >
> >   was perusing the UAPI content in the kernel source, prepping for a
> > course i'm teaching tomorrow, and just noticed there is a header file:
> >
> >   $ find . -name mtd-user.h
> >   ./include/uapi/mtd/mtd-user.h
> >   $
> >
> > which appears to be *entirely* unreferenced in the entire kernel code
> > base (outside of the Kbuild file that refers to it):
> >
> >   $ grep -r 'mtd-user.h' *
> >   include/uapi/mtd/Kbuild:header-y += mtd-user.h
> >   $
>
> Someone has a patch to remove msr-index.h which, in a similar way,
> only ended up being exported because of a Kbuild rule.  So I'd
> imagine this file would be a candidate for a similar patch.
> https://lkml.org/lkml/2017/2/2/228.

  i'm not sure we're talking about the same situations here.

  in the case you mention, it appears that msr-index.h could be
dropped from exporting since (i'm assuming) nothing in user space
needs it. but it's still included from several places in kernel space:

$ grep -r "msr-index.h" *
arch/x86/realmode/rm/wakeup_asm.S:#include <asm/msr-index.h>
arch/x86/realmode/rm/reboot.S:#include <asm/msr-index.h>
arch/x86/xen/enlighten.c:#include <asm/msr-index.h>
arch/x86/boot/cpuflags.c:#include <asm/msr-index.h>
arch/x86/boot/cpucheck.c:#include <asm/msr-index.h>
arch/x86/kernel/verify_cpu.S:#include <asm/msr-index.h>
arch/x86/kernel/head_32.S:#include <asm/msr-index.h>
... etc etc ...
$

so, in this case, that header file is still useful in kernel space and
must be kept.

  in the case *i* mention (mtd-user.h), it appears that that header
file *does* need to exist in user space, but it is totally unused in
kernel space:

$ grep -r "mtd-user.h" *
include/uapi/mtd/Kbuild:header-y += mtd-user.h
$

in cases like this, is it really appropriate for that header file to
exist in kernel space when nothing in kernel space includes it?

  yes, i appreciate the convenience of having all those mtd-related
header files defined in one place; i'm just curious about what the
standard practice is for kernel space defining header files it has no
need for.

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================