Thanks Vlado for your response.
But still its not clear to me how the first operand of the % is int32_t ?
(For the line 202 of
https://github.com/tinyos/tinyos-main/blob/master/tos/lib/ftsp/TimeSyncP.nc)
localAverageRest += (table[i].localTime - newLocalAverage) % tableEntries;
table[i].localTime ---> uint32_t
newLocalAverage ----> uint32_t
tableEntries ---> uint8_t
And therefore localAverageRest will hold only positive values.
This can lead to an error in the calculation of newSkew (through lines
207,213,216,221 of the above mentioned TimeSyncP.nc)
Please correct me if something is missed.
Thanks
Jobish John
On Wed, Oct 4, 2017 at 9:05 PM, Vlado Handziski <handzi...@tkn.tu-berlin.de>
wrote:
> It will not hurt, but I don't see why it is strictly necessary. The first
> operand of the % (more a rest than a modulo) operation is already int32_t,
> the second operand is a uint_8, the results always has the same sign as the
> dividend, as mentioned in the comments.
>
> --Vlado
>
> On Wed, Oct 4, 2017 at 12:31 AM Eric Decker <cire...@gmail.com> wrote:
>
>>
>> I think you are right just looking at the math.
>>
>> However, you will have to test it and watch how the algorithm behaves to
>> make sure it is functioning correctly.
>>
>> I'm not familar enough with the algorithm nor the underlying principles
>> to eye ball it and get it right.
>>
>> If you do figure it out please document it better.
>>
>> thanks,
>>
>> eric
>>
>>
>> On Mon, Oct 2, 2017 at 12:55 AM, Jobish John <jobishjohn.m...@gmail.com>
>> wrote:
>>
>>> Dear all,
>>>
>>> While trying to understand the TinyOS implementation of FTSP, in one of
>>> the function, “calculateConversion()” from (TimeSyncP.nc), I have a
>>> small doubt in the below part,
>>>
>>> ----------------------------------------------------------
>>> while( ++i < MAX_ENTRIES )
>>> if( table[i].state == ENTRY_FULL )
>>> { /*
>>> This only works because C ISO 1999 defines the signe for
>>> modulo the same as for the Dividend!
>>> */
>>> localSum += (int32_t)(table[i].localTime -
>>> newLocalAverage) / tableEntries;
>>> localAverageRest += (table[i].localTime -
>>> newLocalAverage) % tableEntries;
>>> offsetSum += (int32_t)(table[i].timeOffset -
>>> newOffsetAverage) / tableEntries;
>>> offsetAverageRest += (table[i].timeOffset -
>>> newOffsetAverage) % tableEntries;
>>> }
>>> -----------------------------------
>>> I think we need to add a typecast (int32_t) for the localAverageRest
>>> calculation line also, That is, something like
>>> localAverageRest += (int32_t) (table[i].localTime - newLocalAverage) %
>>> tableEntries;
>>>
>>> Kindly suggest..
>>>
>>> Thanks for your valuable time
>>>
>>> Regards
>>> Jobish John
>>> IIT Bombay
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Tinyos-help mailing list
>>> Tinyos-help@millennium.berkeley.edu
>>> https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
>>>
>>
>>
>>
>> --
>> Eric B. Decker
>> Senior (over 50 :-) Researcher
>>
>>
_______________________________________________
Tinyos-help mailing list
Tinyos-help@millennium.berkeley.edu
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
