[ntp:questions] NTP version 2 for control message

Harlan Stenn stenn at ntp.org
Thu Apr 28 15:48:10 UTC 2016


How is this not a bug?

H
--
Wei Wei A writes:
> Hi Martin,
> 
> Thank you so much for your analysis, so can we  say the NTPV2 is hardcoded fo
> r Mode 6 packets?
> 
> Actually I had tried to run "ntpversion 4" from ntpq, but this change doesn't
>  help.
> 
> Br,
> Wei Wei
> 
> 
> 
> -----Original Message-----
> From: Martin Burnicki [mailto:martin.burnicki at burnicki.net] 
> Sent: Wednesday, April 27, 2016 7:11 PM
> To: Wei Wei A; questions at lists.ntp.org
> Subject: Re: [ntp:questions] NTP version 2 for control message
> 
> Wei Wei A wrote:
> > Dear all,
> > 
> > I am running nptd of version 4.2.4p8 on my linux machine,  from wireshark t
> raceing I can see NTPV2 is using for control message, and NTPV4 for client/se
> rver message, could anyone tell me the reason selecting Version 2 for control
>  message, and is that configurable to choose Version 4 for control message.
> > 
> > Your reply is higly appreciated!!
> 
> Hm, I've just run ntpq from the 4.2.8p7 package and indeed wireshark
> still reports NTP version 2 in the flags field of the request packet,
> and also ntpd 4.2.8p7 replies with version 2.
> 
> Even RFC-1305 for NTPv3 states
> 
> --------------------------------------------------------------------
> NTP Control Message Format
> 
> [...]
> Version Number (VN): This is a three-bit integer indicating the NTP
> version number, currently three (3).
> --------------------------------------------------------------------
> 
> so this already should have been 3 in NTPv3. Also RFC-5905 for NTP only
> mentions the version number VN as 4 and with a quick look I havent found
> some exception for control packets. So basically this sounds like a bug.
> 
> On the other hand, a quick glance at the source code yields that a
> symbol NTP_OLDVERSION is defined as 1 with comment "oldest credible
> version", and is used throughout the whole software package.
> 
> Specifically, NTP_OLDVERSION+1 is used when handling the mode 6 packets
> used by ntpq, so it looks like it has been implemented intentionally
> this way. Eventually this just means the mode 6 packet format is
> compatible with NTPv2, but this is just a guess, and it's a yet
> undetected mismatch between the RFCs and the implementation.
> 
> Martin
> 
> _______________________________________________
> questions mailing list
> questions at lists.ntp.org
> http://lists.ntp.org/listinfo/questions
> 


More information about the questions mailing list