Date: Tue, 5 Feb 2002 09:23:17 -0800 (PST) From: John Polstra <jdp@polstra.com> To: hackers@freebsd.org Cc: phk@critter.freebsd.dk Subject: Re: A question about timecounters Message-ID: <200202051723.g15HNH603801@vashon.polstra.com> In-Reply-To: <90115.1012928854@critter.freebsd.dk> References: <90115.1012928854@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
In article <90115.1012928854@critter.freebsd.dk>, Poul-Henning Kamp <phk@critter.freebsd.dk> wrote: > In message <200202051706.g15H6pp03714@vashon.polstra.com>, John Polstra writes: > >In article <XFMail.020204234209.jhb@FreeBSD.org>, > >John Baldwin <jhb@FreeBSD.ORG> wrote: > >> > >> > like, "If X is never locked out for longer than Y, this problem > >> > cannot happen." I'm looking for definitions of X and Y. X might be > >> > hardclock() or softclock() or non-interrupt kernel processing. Y > >> > would be some measure of time, probably a function of HZ and/or the > >> > timecounter frequency. > >> > >> X is hardclock I think, since hardclock() calls tc_windup(). > > > >That makes sense, but on the other hand hardclock seems unlikely to be > >delayed by much. The only thing that can block hardclock is another > >hardclock, an splclock, or an splhigh. And, maybe, splstatclock. I'm > >talking about -stable here, which is where I'm doing my experiments. > > Try swapping so you use the RTC for hardclock & statclock. > > Let the i8254 run with 65536 divisor and do only timecounter service. > > That would be a very interresting experiment. Agreed. But in the cases I'm worrying about right now, the timecounter is the TSC. John -- John Polstra John D. Polstra & Co., Inc. Seattle, Washington USA "Disappointment is a good sign of basic intelligence." -- Chögyam Trungpa To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200202051723.g15HNH603801>