Re: [avr-chat] Missed Optimisation ?

[bob]

On 01/03/11 16:03, Erik Christiansen wrote:

> Errrrr ... Bob, if "result" is in RAM, and so can't be changed by
> hardware, and interrupts have not been reenabled (we're in an ISR),
> then how can "result" be volatile?

In this case, it could be made non-volatile as it's (now) only read
outside of the interrupt, but it wasn't when I started this and noticed
the code :-)

> If memory serves me, here "volatile" is nearly as good as beer, for
> starting a discussion. :-)

Have to disagree, BEER IS FAR MORE IMPORTANT, especially as it's fast
approaching 'Fasching' (some people would say 'carneval', but that's
more northern), here in the Black Forest.

>> It still seems to me that cheating on a part of the code, but not the
>> other part is a bit screwy :-)
> 
> We have too few maintainers for the avr port, and the rest of us are
> grateful to them, but not enough to compete for the job of making a good
> tool "perfect". Oh, and I wonder if there are more little corner cases
> than we would like to tackle.

Fully agree

> It doesn't do any harm to keep one's assembler skills honed, and the
> ISRs are usually short enough to be easily done, barring surprises. 

Ditto

Cheers Erik,

Bob