Subject: Re: re-reading /etc/resolv.conf on change
To: NetBSD Userlevel Technical Discussion List <tech-userlevel@NetBSD.ORG>
From: Greg A. Woods <>
List: tech-userlevel
Date: 01/08/2004 14:58:32
[ On Tuesday, January 6, 2004 at 20:05:19 (-0500), der Mouse wrote: ]
> Subject: Re: re-reading /etc/resolv.conf on change
> (Greg is, however, wrong, if "second" is taken to mean real-world
> second and clock changes are permitted.)

I think the special theory of Unix time relativity applies:  the
observer, being a process running in a system, and the file
last-modified timestamp, also controlled by the same system, cannot
observe or record the real-world clock (and are not even aware of such a
concept), nor can anything inside the system really be aware of clock
skew against real-world time (with the exception of assuming any still
incomplete adjustment initiated by some previous adjtime() call or
equivalent is affecting system time incongruously to real time; and with
the exception of having access to some device connected to an external
real-world clock that can be trusted to be in synchronisation with
real-world time), at least not given the current way such things are
implemented. :-)

(of course a process may be able guess that some gross (> 1 sec) jumps
of the system clock have occurred if it makes certain assumptions about
a file's timestamps and the ability of the system to reliably wake it up
from a sleep() :-)

						Greg A. Woods

+1 416 218-0098                  VE3TCP            RoboHack <>
Planix, Inc. <>          Secrets of the Weird <>