Provide for I/O space access on NetBSD.
Signed-off-by: Jonathan Kollasch jakllsch@kollasch.net ---
2010/10/17 Jonathan A. Kollasch jakllsch@kollasch.net:
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.
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
on 17/10/2010 17:43 Peter Lemenkov said the following:
2010/10/17 Jonathan A. Kollasch jakllsch@kollasch.net:
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/
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 jakllsch@kollasch.net
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
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 jakllsch@kollasch.net
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
On Mon, Oct 18, 2010 at 6:42 AM, Jonathan A. Kollasch jakllsch@kollasch.net 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
Jonathan A. Kollasch wrote:
Provide for I/O space access on NetBSD.
Signed-off-by: Jonathan Kollasch jakllsch@kollasch.net
+++ 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 peter@stuge.se
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 jakllsch@kollasch.net
+++ 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 peter@stuge.se
r5982
But s/__inline/&__/g per comments from Uwe.
-
Andriy Gapon -
Jonathan A. Kollasch -
Peter Lemenkov -
Peter Stuge -
ron minnich -
Warren Turkal