From owner-freebsd-current Thu Jul 16 01:53:51 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id BAA11927 for freebsd-current-outgoing; Thu, 16 Jul 1998 01:53:51 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.15.68.22]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id BAA11920 for ; Thu, 16 Jul 1998 01:53:48 -0700 (PDT) (envelope-from bde@godzilla.zeta.org.au) Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.7/8.8.7) id SAA31913; Thu, 16 Jul 1998 18:53:34 +1000 Date: Thu, 16 Jul 1998 18:53:34 +1000 From: Bruce Evans Message-Id: <199807160853.SAA31913@godzilla.zeta.org.au> To: jak@cetlink.net, phk@critter.freebsd.dk Subject: Re: tickadj -t not changing tick Cc: bde@zeta.org.au, current@FreeBSD.ORG Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG >>You want to fiddle kern.timecounter.frequency (units: Hz) or >>kern.timecounter.adjustment (Units: PPM/2^16) > >Judging from what Bruce said, it seems that kern.timecounter.frequency >is derived from the machdep.*_freq variables, so it would be better to >change them and let the kernel propagate the changes. It would be nicest to change the machine-independent variables and let the kernel propagate the changes to the machine-dependent variables and hardware. However the kernel makes no attempt to propagate the changes. OTOH, when you change the m-d variables, the kernel attempts to propagate the changes, but it doesn't succeed (it sets the active frequency, possibly causing a glitch due to the non-atomic update, and the change gets blown away at the next clock interrupt). >And if the appropriate machdep.*_freq variable is set precisely, is >there any need for kern.timecounter.adjustment? Only in practice. Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message