Le 11/5/2007, "Jan Kiszka" <[EMAIL PROTECTED]> a écrit:
>CHABAL David wrote:
>>>> In additional:
>>>> 1. You should use 921600 value instead of 912.600.
>> The "." was here for reading facility only, I consider it as an integer
>> in my computations.
>The point was not the point, but the flipped digits.
>>> (921600 + (115200 div 2)) div 115200 = 8
>>> So back to square #1: misconfiguration? David, please check with the
>>> correct baud base again if you still see relevant divider variations.
>> moxa: 921600/115200 = 8 (not 7, I apologize for this mistake but we
>> guess here that the baud base is always 921600 and according to the
>> mxser file it can be 230400 for some cards)
>(230400+57600)/115200=2,5 => 2
Oupsss, I'm beaten :-(
The only case with a different value that I found is for 1200:
(230400+1200)/1200 = 193
230400/1200 = 192
I think it's not relevant for this speed.
>> xeno: (921600+57600)/115200 = 8.5 rounded to 8
>So we have the same results for both formulas and I don't see the need
>for a patch anymore. Do you? Otherwise I would just commit the
>div-by-zero fix to trunk now and keep the rest as is.
Xenomai-core mailing list