[patch] superiotool NetBSD support

[Jonathan A. Kollasch]

Provide for I/O space access on NetBSD.

Signed-off-by: Jonathan Kollasch 
---

[Peter Lemenkov]

2010/10/17 Jonathan A. Kollasch :
> Provide for I/O space access on NetBSD.

It seems that it's time for flashrom maintainers to reconsider their
opinion regarding autotools since the complexity and number of nesting
levels of ifdefs within flashrom sources  becomes more and more
difficult to read and understand.

[ron minnich]

Yes, I suppose you can do this with autotools. I've never seen
autotools as necessary or sufficient. Not necessary, because I use far
more complex projects (see swtch.com/plan9ports, which is a complete
Plan 9 operating system including graphics; or the Go language from
Google) that somehow manage to get along without all the overhead that
comes with autotools. Not sufficient, because I've lost track of the
number of times autotools have failed me on one system or another.

I think this #ifdef might be done a bit better but its existence is
not a strong case for autotools.

ron

[Andriy Gapon]

on 17/10/2010 17:43 Peter Lemenkov said the following:
> 2010/10/17 Jonathan A. Kollasch :
> > Provide for I/O space access on NetBSD.
>
> It seems that it's time for flashrom maintainers to reconsider their
> opinion regarding autotools since the complexity and number of nesting
> levels of ifdefs within flashrom sources  becomes more and more
> difficult to read and understand.

BTW:
http://sourceforge.net/projects/mk-configure/

[Warren Turkal]

On Sunday, October 17, 2010 07:39:06 am Jonathan A. Kollasch wrote:
> Provide for I/O space access on NetBSD.
> Signed-off-by: Jonathan Kollasch

It might be better to use "inline" instead of "__inline" and define a macro to 
use the real name you want on certain platorms since c99 reserves the keyword 
"inline" for this purpose.

Having said that, do those functions really need to be inlined?

Thanks,
wt

[Jonathan A. Kollasch]

On Sun, Oct 17, 2010 at 10:51:06PM -0700, Warren Turkal wrote:
> On Sunday, October 17, 2010 07:39:06 am Jonathan A. Kollasch wrote:
> > Provide for I/O space access on NetBSD.
> > Signed-off-by: Jonathan Kollasch
>
> It might be better to use "inline" instead of "__inline" and define a macro to 
> use the real name you want on certain platorms since c99 reserves the keyword 
> "inline" for this purpose.

Well, the Makefile gives the compiler -ansi ...
> Having said that, do those functions really need to be inlined?

Maybe, they aren't used that much, and may as well flow right into
the code IMO.

Jonathan Kollasch

[Warren Turkal]

On Mon, Oct 18, 2010 at 6:42 AM, Jonathan A. Kollasch
 wrote:
> On Sun, Oct 17, 2010 at 10:51:06PM -0700, Warren Turkal wrote:
> > It might be better to use "inline" instead of "__inline" and define a macro to
> > use the real name you want on certain platorms since c99 reserves the keyword
> > "inline" for this purpose.
>
> Well, the Makefile gives the compiler -ansi ...

The Makefile can be changed.
> > Having said that, do those functions really need to be inlined?
>
> Maybe, they aren't used that much, and may as well flow right into
> the code IMO.

Two counters:
* GCC can inline functions that don't declare themselves inline, and
it can refuse to inline one that use the __inline attribute. I think
it'd be better to rely on compiler optimizations than to explicitly
state that it's inline.
* Superiotool isn't some long running process, so an optimization that
makes the code less readable is not really useful. I believe that it
makes the code less readable by embedding functional code (i.e. not
just prototypes and macros) in a header file.'

Thanks,
wt

[Peter Stuge]

Jonathan A. Kollasch wrote:
> Provide for I/O space access on NetBSD.
> Signed-off-by: Jonathan Kollasch
> +++ Makefile	(working copy)

Please keep in mind to generate patches from the top-level checkout
directory so that it's always clear which files are being touched.
Thanks!

Acked-by: Peter Stuge

[Jonathan A. Kollasch]

On Tue, Oct 19, 2010 at 03:21:18AM +0200, Peter Stuge wrote:
> Jonathan A. Kollasch wrote:
> > Provide for I/O space access on NetBSD.
> > Signed-off-by: Jonathan Kollasch
> > +++ Makefile	(working copy)
>
> Please keep in mind to generate patches from the top-level checkout
> directory so that it's always clear which files are being touched.
> Thanks!
>
> Acked-by: Peter Stuge

r5982

But s/__inline/&__/g  per comments from Uwe.