[ntp:hackers] Why are we going down this road? Question on listen-on, query-on, -I
stenn at ntp.org
Thu May 28 08:06:12 UTC 2009
There are a number of places folks would like to have better control
over things. For example, modem configuration strungs, baud rate
changes, the number of addresses to accept for a hostname, better
logging and debug control, memory locking control, configurable audio
refclock parameters, allowing "fudge" for non-refclocks, etc.
The recent discussion about "listen-on" and several long-standing bug
reports point to the need to allow more than the current "open
everything, and open-read-drop based on 'restrict' parameters".
Once we get a better handle on the things we're looking at, we'll be
able to see if we want to extend the current parser with more keywords,
or look at using a config file syntax that is more like the one BIND
uses (for example). Even if we change the syntax a bit, we may still be
able to use the latest parser framework.
It's just something to look at and evaluate.
> Please explain what you mean by "rewrie of the ntp.conf file
> format/syntax". We spent considerable effort and not a little funds over
> two years developing a syntax-directed, table-defined parser. You wil
> need to serisously defend you problems with that.
> Harlan Stenn wrote:
> >From another angle, I'd really like to see if we can come up with a
> >useful fix for "don't open IP xxx" for 4.2.6, as it will solve several
> >long-standing bug reports.
> >I also want to see a rewrite of the ntp.conf file format/syntax.
> >I believe that addressing "open-read, open-read-drop, don't-open" will
> >be a Learning Experience, and if we can get this going for 4.2.6 we will
> >be able to have have a *much* better solution with an ntp.conf rewrite
> >for 4.3.0.
> hackers mailing list
> hackers at lists.ntp.org
More information about the hackers