[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [avr-libc-dev] <avr/sfr_defs.h> documented
From: |
Joerg Wunsch |
Subject: |
Re: [avr-libc-dev] <avr/sfr_defs.h> documented |
Date: |
Thu, 22 Aug 2002 18:22:14 +0200 |
User-agent: |
Mutt/1.2.5i |
As Theodore A. Roth wrote:
> :) My only concern with it was whether all this is really applicable to
> :) assembler programmers at all. I. e., are there people who would
> :) really like to tweak all those knobs?
>
> That's why it's a sub-page. You can skip it if they wish, but it's there
> for the adventurous.
OK, convinced. ;-)
> As I said in a previous email, I am mostly neutral on this as long as it
> doesn't break user code by making them have to explicitly include the
> compat.h file.
See my other mail, we broke it anyway. If we don't want to provide
the backwards compatibility, we could as well drop outp()/inp()/BV()
completely. We've already dropped so many other things that slowly
became documented and used in the past.
> :) Has there actually been an outw() macro? I thought there was only
> :) __outw() (which is gone now).
>
> Accroding to the source, there seems to be an outw() macro.
Yep, i just installed an old package to review it. There was a subtle
difference between outw() and __outw() though. __outw() was always
using OUT instructions, while outw() decided whether to use __outw()
or memory-mapped IO instead. If i understand it right, this is all a
decision of the compiler (and optimzer) in the current gcc
implementation.
But OK, the changed argument order for outb() and outw() is already
enough requirement for the users to adapt their programs. IMHO we
could drop all other backward-compatibility macros then.
--
J"org Wunsch Unix support engineer
address@hidden http://www.interface-systems.de/~j/