Re: [avr-gcc-list] What GCC bug currently effect the AVR port of GCC?

[Brian Dean]

On Mon, Feb 28, 2005 at 11:46:27PM +0100, Joerg Wunsch wrote:

> Well, it claims GCC 3.3 had been fixed for the "volatile" case.
> Strange, why has that fix been undone then?

I thought I remembered you telling me that Marek had either committed
your patch or developed an improved version.  That's why I started the
thread entited "Old bug ...".  I thought it had be previously fixed
and later reverted.

But since I got my G5 (no nice FreeBSD ports tree), I've been building
gcc the old fashioned way as opposed to using the much more civilized
FreeBSD port and your excellent port of it :-).  I was willing to
believe that I only got the "correct" behaviour because I was using
your patched FreeBSD port and when I started using a Mac OS X system
with my own built avr-gcc binary, I didn't have your patch.

But are you now thinking that this was once fixed but has now been
reverted, either accidentally or on purpose?  If accidentally, then it
shouldn't be too much trouble to get it back in.  If it was removed on
purpose, I can't fathom the reasoning, since it's clearly a bug on the
AVR architecture to write 16-bit values low-byte first for memory
mapped i/o registers.

-Brian
-- 
Brian Dean
BDMICRO - ATmega128 Based MAVRIC Controllers
http://www.bdmicro.com/