Re: [psas-avionics] Problems With The Flight Computer Log File

2013-06-15 Thread James Perkins
Something I learned this year at work... if you use syslog(2) for logging
it can block for long periods if the log daemon stalls in reading the
socket (mine stalls to write to flash for example). For my work application
I now log strings to a memory circular buffer and then have an asynchronous
thread for each of my application processes which pulls messages out of the
buffer and calls syslog.
On Jun 14, 2013 10:42 PM, "Clark Wachsmuth" 
wrote:

> My bad for sure on iii). I was putting a line break on the end of each
> packet so it made things more human readable for me. I've taken that out
> and pushed it the the repo.
>
>
> On Fri, Jun 14, 2013 at 5:34 PM, Nathan Bergey wrote:
>
>> The good news on Tuesday is that we ran the FCF on the actual flight
>> computer, and even recorded a log file!
>>
>> I finally got to sit down today and try and read the data back out and
>> found this:
>>
>> I didn't see any packet headers (only message headers), but that's
>> fine, I don't know where those get added, maybe only as it's dumped to
>> WiFi.
>>
>> Problems:
>>
>> 1.) The data all looked little-endian and I think we decided that
>> packets should be converted to network (big) endian before being sent.
>> So presumably we should log them this way too.
>>
>> B.) The 6 byte time stamps were always all zeros.
>>
>> iii.) The packet lengths seemed to all be exactly 1 byte longer than
>> they should be. For instance the MPL header looked like 'MPL3', 6 byte
>> timetamps (0's), 2 byte length (8), and then *9* bytes of data. It's
>> looked the same for the others as far as I could tell.
>>
>> So, it's close to a valid logfile. Clearly there is some data there
>> but I can't quite read it with the tools we wrote because of above
>> issues. Can someone double check the message making code on the FC and
>> see if what I think the log file looks like makes sense with what was
>> written?
>>
>> -Nathan
>>
>> ___
>> psas-avionics mailing list
>> psas-avionics@lists.psas.pdx.edu
>> http://lists.psas.pdx.edu/mailman/listinfo/psas-avionics
>>
>
>
> ___
> psas-avionics mailing list
> psas-avionics@lists.psas.pdx.edu
> http://lists.psas.pdx.edu/mailman/listinfo/psas-avionics
>
>
___
psas-avionics mailing list
psas-avionics@lists.psas.pdx.edu
http://lists.psas.pdx.edu/mailman/listinfo/psas-avionics


Re: [psas-avionics] Problems With The Flight Computer Log File

2013-06-14 Thread Clark Wachsmuth
My bad for sure on iii). I was putting a line break on the end of each
packet so it made things more human readable for me. I've taken that out
and pushed it the the repo.


On Fri, Jun 14, 2013 at 5:34 PM, Nathan Bergey  wrote:

> The good news on Tuesday is that we ran the FCF on the actual flight
> computer, and even recorded a log file!
>
> I finally got to sit down today and try and read the data back out and
> found this:
>
> I didn't see any packet headers (only message headers), but that's
> fine, I don't know where those get added, maybe only as it's dumped to
> WiFi.
>
> Problems:
>
> 1.) The data all looked little-endian and I think we decided that
> packets should be converted to network (big) endian before being sent.
> So presumably we should log them this way too.
>
> B.) The 6 byte time stamps were always all zeros.
>
> iii.) The packet lengths seemed to all be exactly 1 byte longer than
> they should be. For instance the MPL header looked like 'MPL3', 6 byte
> timetamps (0's), 2 byte length (8), and then *9* bytes of data. It's
> looked the same for the others as far as I could tell.
>
> So, it's close to a valid logfile. Clearly there is some data there
> but I can't quite read it with the tools we wrote because of above
> issues. Can someone double check the message making code on the FC and
> see if what I think the log file looks like makes sense with what was
> written?
>
> -Nathan
>
> ___
> psas-avionics mailing list
> psas-avionics@lists.psas.pdx.edu
> http://lists.psas.pdx.edu/mailman/listinfo/psas-avionics
>
___
psas-avionics mailing list
psas-avionics@lists.psas.pdx.edu
http://lists.psas.pdx.edu/mailman/listinfo/psas-avionics