Re: [ntp:questions] NTP, GPSD PPS
On 10/12/2014 12:39, Miroslav Lichvar wrote: On Wed, Dec 10, 2014 at 11:50:22AM +, David Taylor wrote: With -D 4 I get a list of devices ending with PPS, but presumably that is not the same as KPPS? In gpsd the PPS without K is the userspace timestamping. With kernel timestamping the log looks like this: gpsd:PROG: PPS edge: 1, cycle: 100 uSec, duration: 78 uSec @ 1418214654.07216 gpsd:INFO: PPS hooks called with accepted 1418214653.99223 offset 0.00777 gpsd:PROG: PPS edge accepted 1418214653.99223 offset 0.00777 gpsd:PROG: KPPS assert 1418214653.99223, sequence: 73 - clear 1418214654.20573, sequence: 73 gpsd:PROG: KPPS data: using clear gpsd:PROG: KPPS cycle: 99 uSec, duration: 21 uSec @ 1418214654.20573 I did try an apt-get first to update gpsd but it seems I have the most recent available. It seems I have 3.6. Do I need a development version or...? The kernel PPS support was added in 3.0 or so, but gpsd needs to be compiled with the timepps.h header, similarly to ntpd for the ATOM and other drivers. Also, some gpsd versions had bugs in the PPS/KPPS support and I'm not sure if 3.6 was a good or bad. The latest version - 3.11 is working well for me, 3.10 was not. Many thanks, Miroslav. For some reason, gpsd seems to be stuck at a lower release for the Raspberry Pi version of Debian, so I will have a go at recompiling it once I can find the right instructions! That's if there's no development version I can get with apt-get, and I haven't found out how to do that yet. Still learning! -- Cheers, David Web: http://www.satsignal.eu ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTP PPS, part 2 ;)
Harlan Stenn st...@ntp.org wrote: Rob writes: Harlan Stenn st...@ntp.org wrote: Rob writes: Harlan Stenn st...@ntp.org wrote: If you disgree and think NTP should provide the file all the time, then: - how do you propose we find out if the underlying API is really provided in the currently-running kernel? The source of the includefile does absolutely nothing in the ways of solving that problem! If the file isn't there we don't go looking for the API that isn't there, either. Or am I missing something? The file is only used at build time. It tells absolutely nothing about the kernel configuration, certainly not in the system the binary is running on. You and I have completely different understandings about how APIs work and what this header file is used for. So you want *us* to add kernel-specific files to live along side include/timepps-{SCO,Solaris,SunOS}.h, except you want *us* to deal with tracking any changes caused by kernel updates? It's interesting enough that we have to do this for Windows. I am not commenting on the whoe provides what but on your claim that using an available timepps.h would do anything to detect if the PPS API is available on the system. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTP PPS, part 2 ;)
Harlan Stenn st...@ntp.org wrote: Paul writes: --001a11c12566ef4fbd050a04ed7c Content-Type: text/plain; charset=ISO-8859-1 On Dec 12, 2014 12:39 AM, Harlan Stenn st...@ntp.org wrote: It's an OS-specific file that should be provided by the OS if the underlying API exists. To repeat what I reminded you of last time. Linux *doesn't* have the API. The macros in timepps provide the RFC compliant API. The NTP developers should stop depending on the pps-tools maintainer to provide the macros and rewrite the module to use the native ioctl interface or ask the downstream maintainers to take on that task. Who wants to do this work? NTF will take it on after it gets funding for developers. I'd love to see that happen sooner rather than later. Oh come on... this is just a matter of copying one file of a few KB into the include directory of the ntp package and you are done. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTP PPS, part 2 ;)
Harlan Stenn st...@ntp.org wrote: Rob writes: Harlan Stenn st...@ntp.org wrote: Martin Burnicki writes: IMO the best approach would be to detect this at runtime. That means we'd need a header file... If I'm not mistaken (and it's getting late for me), if the header file is missing we don't expect the API. If the header file is present we expect it to do the right thing and even then we check error returns from the API. But the problem is the file is normally not there, even when the API is. No, the problem is that it's not normally there on *linux* boxes where the file is maintained in a separate package. It *is* provided by the other OSes that implement this API. Thos OSes probably have no package management and/or always install everything even when the user has not asked for it. Again, the solution is simple: building a correct ntpd depends on the availability of the pps-tools package on the build system. It should be added to the build dependencies to solve the problem. Current build dependencies on Debian are: debhelper (= 4.2.12), libreadline-dev, lynx, libcap-dev, libedit-dev, libperl-dev, libssl-dev (= 0.9), libsnmp-dev, quilt (= 0.40), libevent-dev Add pps-tools to that list and the problem is over, finito. The only problem we face is that people don't want to type 'A, pps-toolsESCZZ' and finish a long-running discussion. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTP, GPSD PPS
On Sat, Dec 13, 2014 at 2:52 AM, David Taylor david-tay...@blueyonder.co.uk.invalid wrote: For some reason, gpsd seems to be stuck at a lower release for the Raspberry Pi version of Debian, so I will have a go at recompiling it once I can find the right instructions! That's if there's no development version I can get with apt-get, and I haven't found out how to do that yet. Still learning! If you don't want to cross-compile you need gcc, libncurses5-dev and scons. It's slow but simple. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTP PPS, part 2 ;)
On Sat, Dec 13, 2014 at 5:01 AM, Rob nom...@example.com wrote: Current build dependencies on Debian are: That's better said on my install of Debian. I wouldn't expect it be the case on all release tracks and it doesn't help Ubuntu. Of course for an S1 operator the fact that this approach means building 4.2.6p5 or even earlier is another insufficiency. I don't understand why you continue to suggest using a stale source tree for building S1 servers. It's no harder to build from the developer source. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTP PPS, part 2 ;)
Paul tik-...@bodosom.net wrote: On Sat, Dec 13, 2014 at 5:01 AM, Rob nom...@example.com wrote: Current build dependencies on Debian are: That's better said on my install of Debian. I wouldn't expect it be the case on all release tracks and it doesn't help Ubuntu. Of course for an S1 operator the fact that this approach means building 4.2.6p5 or even earlier is another insufficiency. I don't understand why you continue to suggest using a stale source tree for building S1 servers. It's no harder to build from the developer source. You know what? On the ntp-dev package for Debian THE BUILD DEPENDENCIES ARE INCORRECT AS WELL!! Those kind of things are hard, very hard, to get fixed. In fact on OpenSUSE it is the same thing. It has nothing to do with the ntpd source version, it is just the bickering about who provides what and who puts what in their package description. Apparently when you want to run a server PPS, you have to build the binary yourself. For no real reason, of course. Fortunately there is gpsd. At least those people know that you need to write a selfcontained package with directions and examples when you want things to be right. (another failure of the ntpd project: not providing a usable default config file, so distributors have to put one together and causing config mistakes because it is not their field of knowledge) ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
[ntp:questions] pool.ntp.org and authentication
Hi all I was wondering if it makes sense to set up Autokey authentication on a client for when it wants to sync time from *.pool.npt.org. My goal is to encrypt communication between client and server and to make sure the server is really the one it claims to be. Can this be even done with pools? ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTP PPS, part 2 ;)
On Sat, Dec 13, 2014 at 3:16 PM, Rob nom...@example.com wrote: You know what? On the ntp-dev package for Debian THE BUILD DEPENDENCIES ARE INCORRECT AS WELL!! This is an example of what NTF doesn't want to deal with. My instance of Wheezy doesn't have ntp-dev. Fortunately there is gpsd. You do realize gpsd pps support requires timepps which isn't in the 3.9 or 3.11 tarballs (those are the ones I have at hand). ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTP PPS, part 2 ;)
Rob writes: Harlan Stenn st...@ntp.org wrote: Paul writes: --001a11c12566ef4fbd050a04ed7c Content-Type: text/plain; charset=ISO-8859-1 On Dec 12, 2014 12:39 AM, Harlan Stenn st...@ntp.org wrote: It's an OS-specific file that should be provided by the OS if the underlying API exists. To repeat what I reminded you of last time. Linux *doesn't* have the API. The macros in timepps provide the RFC compliant API. The NTP developers should stop depending on the pps-tools maintainer to provide the macros and rewrite the module to use the native ioctl interface or ask the downstream maintainers to take on that task. Who wants to do this work? NTF will take it on after it gets funding for developers. I'd love to see that happen sooner rather than later. Oh come on... this is just a matter of copying one file of a few KB into the include directory of the ntp package and you are done. You have clearly never done releng work. H ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] pool.ntp.org and authentication
d_anderson writes: Hi all I was wondering if it makes sense to set up Autokey authentication on a client for when it wants to sync time from *.pool.npt.org. My goal is to encrypt communication between client and server and to make sure the server is really the one it claims to be. Can this be even done with pools? Not with the current technology. First, autokey is about to become deprecated in favor of NTS - Network Time Security: https://tools.ietf.org/html/draft-ietf-ntp-network-time-security-05 If that wasn't the case, autokey (which was designed a long time ago) needs the server to have a unique key. For pool servers, every pool server would have to share the same private key. That would make the security provided almost nonexistent. If we changed the protocol to use some other mechanism to get the server's key (probably based on the IP) we'd need to change the autokey protocol. That would not appear to be a worthwhile exercise given that we intend to deprecate autokey in favor of NTS soon. H ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] Number of Stratum 1 Stratum 2 Peers
On 2014-12-12 03:25, Harlan Stenn wrote: It's pretty easy to download and install a leapsecond file, and ntpd will pay attention to that... Not that easy - unless you are one of the lucky few to have encrypted access to a NIST source, when it may be automatic. You have to use a NIST server, as no other sources provide access to the NIST leapseconds file, find one where FTP access is available and/or works reliably from your system, schedule a download every six months, check the signature, and if all goes well, replace your current file. They also use their own weird approach to checking the file signature from the last century, and source code to build to do so, rather than standard approaches built into utilities available for and on modern systems. You also have to specify in ntp.conf where the leapseconds file is stored, whereas most other external configuration information can be passed on the ntpd command line. It would be interesting to know what percentage of the pool servers even use a leapseconds file, and how many of those have a valid copy. I am certain that very few clients use a leapseconds file. OTOH the timezone/zoneinfo package uses its own leapseconds file (for right time - now zoneinfo-leaps), and distributes that and the original, a script that checks and converts it to their own format, and utilities that use it. -- Take care. Thanks, Brian Inglis ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] Number of Stratum 1 Stratum 2 Peers
Brian Inglis writes: On 2014-12-12 03:25, Harlan Stenn wrote: It's pretty easy to download and install a leapsecond file, and ntpd will pay attention to that... Not that easy - unless you are one of the lucky few to have encrypted access to a NIST source, when it may be automatic. http://www.ietf.org/timezones/data/leap-seconds.list H ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] Number of Stratum 1 Stratum 2 Peers
On 14/12/14 03:28, Harlan Stenn wrote: Not that easy - unless you are one of the lucky few to have encrypted access to a NIST source, when it may be automatic. http://www.ietf.org/timezones/data/leap-seconds.list Added to the Wiki at http://support.ntp.org/bin/view/Support/ConfiguringNTP The IETF also serve their content over SSL if anyone thinks this increases the level of trust one can have in that content. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions