[KJ] centralizing all of the types.h content?

[KJ] centralizing all of the types.h content?

[Robert P. J. Day]

  is there any reason that most of the content of the various
include/asm-*/types.h files couldn't be centralized, perhaps in
include/linux/types.h?

  certainly, a lot of that content looks absolutely identical across
all architectures.

rday

-- 
====================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry
Waterloo, Ontario, CANADA

http://fsdev.net/wiki/index.php?title=Main_Page
====================================
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/kernel-janitors

Re: [KJ] centralizing all of the types.h content?

[Alexey Dobriyan]

On 5/15/07, Robert P. J. Day <rpjday@mindspring.com> wrote:
>   is there any reason that most of the content of the various
> include/asm-*/types.h files couldn't be centralized, perhaps in
> include/linux/types.h?

Yes! Lots of ifdefs, because files aren't quite identical.

>   certainly, a lot of that content looks absolutely identical across
> all architectures.
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/kernel-janitors

Re: [KJ] centralizing all of the types.h content?

[Robert P. J. Day]

On Tue, 15 May 2007, Alexey Dobriyan wrote:

> On 5/15/07, Robert P. J. Day <rpjday@mindspring.com> wrote:
> >   is there any reason that most of the content of the various
> > include/asm-*/types.h files couldn't be centralized, perhaps in
> > include/linux/types.h?
>
> Yes! Lots of ifdefs, because files aren't quite identical.

oh, i realize that there are differences, but isn't there a lot of
duplicated content across all those files that will always
*necessarily* be identical?

i mean, is there a single architecture that *won't*, say, define the
following typedef:

  typedef __signed__ char __s8;

will there ever be an architecture for which the typedef "__s8"
*won't* be a signed char?  and if there might be, it would still be
more efficient to create a general include/asm-generic/types.h that
most people could use, and let some weird and wacky architecture
always override that with its own.

even with all those #ifdef's, it still seems like there's a lot of
common content whose duplication is unnecessary.

rday
-- 
====================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry
Waterloo, Ontario, CANADA

http://fsdev.net/wiki/index.php?title=Main_Page
====================================
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/kernel-janitors