Date: Fri, 10 Jul 2009 08:50:47 +0300 From: Andriy Gapon <avg@freebsd.org> To: Tim Kientzle <kientzle@freebsd.org> Cc: =?ISO-8859-1?Q?Marius_N=FCnnerich?= <marius@nuenneri.ch>, freebsd-current@freebsd.org Subject: Re: dtrace users opinion solicited (timestamps) Message-ID: <4A56D6B7.8050100@freebsd.org> In-Reply-To: <4A56AD55.2010201@freebsd.org> References: <4A562960.3010801@freebsd.org> <b649e5e0907091102h2bbf3799r4f8b840696a9162b@mail.gmail.com> <4A563A57.8090907@freebsd.org> <4A56AD55.2010201@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
on 10/07/2009 05:54 Tim Kientzle said the following: > > Instead of just storing tsc_freq at each frequency change, > you really need: > * Last timestamp just before frequency change (t0) > * new frequency (f) > * TSC value at frequency change (c0) > > Then > t = t0 + (rdtsc() - c0) / f > > Of course, I haven't looked at the code to tell > what your range limitations are. Yes, tracking the TSC frequency changes should allow us to have correct DTrace timecounting. But I am not sure if we really need this (or can have this). Because otherwise I can not see why we have a distinct/specialized DTrace TSC-based timecounting when we already have general purpose TSC timecounting that already works correctly (if I am not mistaken). -- Andriy Gapon
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4A56D6B7.8050100>