In a recent note, David Andrews said: > Date: Thu, 10 Nov 2005 10:28:25 -0500 > > On Thu, 2005-11-10 at 08:18 -0700, Paul Gilmartin wrote: > > In a recent note, David Andrews said: > > > the TOD clock drifts over time (we don't have > > > an ETR) and we adjust the local offset every week to compensate. > > > > > If you're forced to this deceit, would it be better to chisel on > > the value of the Leap Second offset? > > This "deceit" is accomplished by a program that interrogates a NTP > server out on the greater 'net, and then constructs a "SET CLOCK" > command. A poor man's ETR. > If that program were suitably authorized to use storage key 0, it could as well update the leap second offset in the CVT. I believe a doubleword-aligned STM is atomic, so there's no hazard of other processes seeing a partially updated value. There would remain the hazard of other processes seeing ambiguous GMT timestamps in the case of a backward adjustment.
-- gil -- StorageTek INFORMATION made POWERFUL ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
