Subject: Re: lazy mlock?
To: David Laight <>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: tech-kern
Date: 04/16/2002 16:10:30
In message <>David Laight writes
>If it isn't a silly question, why should ntpd 'lock down' any code at all?

To eliminate jitter due to pagefaults in ntpd's code.  ntp's lockdown
of its pages dates back to, oh, ntp v2 on Sunos 4.1.x, Ultrix 4.x, usw.

A well-behaved ntpd sits idle for up to 15-odd minutes between clock
updates. Without the lockdown, it looks like a perfect candidate for
pageout. IIRC, modern ntpds can be configured to send multiple
`chimes' each time the update timer expires.  Discussion of the
tradeoffs belongs on comp.protocols.time.ntp.