[PATCH] Drop an empty byteorder/generic.h from being exported to userspace.

[PATCH] Drop an empty byteorder/generic.h from being exported to userspace.

[Robert P. J. Day]

Since the entire meaningful contents of generic.h is wrapped by the
__KERNEL__ test, drop it from being passed to userspace, and remove
the now superfluous __KERNEL__ test.

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>

---

  is there some value in allowing an effectively empty header file to
be exported to userspace?

 include/linux/byteorder/Kbuild    |    1 -
 include/linux/byteorder/generic.h |    8 --------
 2 files changed, 9 deletions(-)

diff --git a/include/linux/byteorder/Kbuild b/include/linux/byteorder/Kbuild
index 79beddd..a50614c 100644
--- a/include/linux/byteorder/Kbuild
+++ b/include/linux/byteorder/Kbuild
@@ -1,5 +1,4 @@
 header-y += big_endian.h
 header-y += little_endian.h

-unifdef-y += generic.h
 unifdef-y += swab.h
diff --git a/include/linux/byteorder/generic.h b/include/linux/byteorder/generic.h
index d377155..0846e6b 100644
--- a/include/linux/byteorder/generic.h
+++ b/include/linux/byteorder/generic.h
@@ -82,12 +82,6 @@
  *
  */

-
-#if defined(__KERNEL__)
-/*
- * inside the kernel, we can use nicknames;
- * outside of it, we must avoid POSIX namespace pollution...
- */
 #define cpu_to_le64 __cpu_to_le64
 #define le64_to_cpu __le64_to_cpu
 #define cpu_to_le32 __cpu_to_le32
@@ -176,6 +170,4 @@ static inline void be64_add_cpu(__be64 *var, u64 val)
 	*var = cpu_to_be64(be64_to_cpu(*var) + val);
 }

-#endif /* KERNEL */
-
 #endif /* _LINUX_BYTEORDER_GENERIC_H */
========================================================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry:
    Have classroom, will lecture.

http://crashcourse.ca                          Waterloo, Ontario, CANADA
========================================================================

Re: [PATCH] Drop an empty byteorder/generic.h from being exported to userspace.

[Jiri Slaby]

On 04/02/2008 03:38 AM, Robert P. J. Day wrote:
> Since the entire meaningful contents of generic.h is wrapped by the
> __KERNEL__ test, drop it from being passed to userspace, and remove
> the now superfluous __KERNEL__ test.
> 
> Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
> 
> ---
> 
>   is there some value in allowing an effectively empty header file to
> be exported to userspace?

Not breaking the userspace? The file is included in non-__KERNEL__ sections from 
byteorder/*_endian.h.

Re: [PATCH] Drop an empty byteorder/generic.h from being exported to userspace.

[Robert P. J. Day]

On Wed, 2 Apr 2008, Jiri Slaby wrote:

> On 04/02/2008 03:38 AM, Robert P. J. Day wrote:
> > Since the entire meaningful contents of generic.h is wrapped by the
> > __KERNEL__ test, drop it from being passed to userspace, and remove
> > the now superfluous __KERNEL__ test.
> >
> > Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
> >
> > ---
> >
> >   is there some value in allowing an effectively empty header file to
> > be exported to userspace?
>
> Not breaking the userspace? The file is included in non-__KERNEL__
> sections from byteorder/*_endian.h.

ok ... and the value of including a completely empty header file is
... ?

rday
--


========================================================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry:
    Have classroom, will lecture.

http://crashcourse.ca                          Waterloo, Ontario, CANADA
========================================================================

Re: [PATCH] Drop an empty byteorder/generic.h from being exported to userspace.

[Jiri Slaby]

On 04/02/2008 01:00 PM, Robert P. J. Day wrote:
> On Wed, 2 Apr 2008, Jiri Slaby wrote:
> 
>> On 04/02/2008 03:38 AM, Robert P. J. Day wrote:
>>> Since the entire meaningful contents of generic.h is wrapped by the
>>> __KERNEL__ test, drop it from being passed to userspace, and remove
>>> the now superfluous __KERNEL__ test.
>>>
>>> Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
>>>
>>> ---
>>>
>>>   is there some value in allowing an effectively empty header file to
>>> be exported to userspace?
>> Not breaking the userspace? The file is included in non-__KERNEL__
>> sections from byteorder/*_endian.h.
> 
> ok ... and the value of including a completely empty header file is
> ... ?

No value, you should just move those includes into the __KERNEL__ part.

Re: [PATCH] Drop an empty byteorder/generic.h from being exported to userspace.

[Robert P. J. Day]

On Wed, 2 Apr 2008, Jiri Slaby wrote:

> On 04/02/2008 01:00 PM, Robert P. J. Day wrote:
> > On Wed, 2 Apr 2008, Jiri Slaby wrote:
> >
> > > On 04/02/2008 03:38 AM, Robert P. J. Day wrote:
> > > > Since the entire meaningful contents of generic.h is wrapped by the
> > > > __KERNEL__ test, drop it from being passed to userspace, and remove
> > > > the now superfluous __KERNEL__ test.
> > > >
> > > > Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
> > > >
> > > > ---
> > > >
> > > >   is there some value in allowing an effectively empty header file to
> > > > be exported to userspace?
> > > Not breaking the userspace? The file is included in non-__KERNEL__
> > > sections from byteorder/*_endian.h.
> >
> > ok ... and the value of including a completely empty header file is
> > ... ?
>
> No value, you should just move those includes into the __KERNEL__ part.

gotcha, that's what i'd concluded after your earlier post, i just
wasn't sure if there was something more subtle happening here.
revised patch coming soon.  thanks.

rday
--

========================================================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry:
    Have classroom, will lecture.

http://crashcourse.ca                          Waterloo, Ontario, CANADA
========================================================================