Date: Mon, 25 Jan 2010 17:58:11 +0100 From: Attilio Rao <attilio@freebsd.org> To: Oliver Fromme <olli@lurza.secnetix.de> Cc: freebsd-current@freebsd.org, Richard Todd <rmtodd@ichotolot.servalan.com> Subject: Re: top(1) + vmstat(8): CPU percentages broken Message-ID: <3bbf2fe11001250858p2516e20y95576eb3cc59764c@mail.gmail.com> In-Reply-To: <201001251605.o0PG5dGn026347@lurza.secnetix.de> References: <3bbf2fe11001250648n4f02e33eu1bf44df5370cd28b@mail.gmail.com> <201001251605.o0PG5dGn026347@lurza.secnetix.de>
next in thread | previous in thread | raw e-mail | index | archive | help
2010/1/25 Oliver Fromme <olli@lurza.secnetix.de>: > Attilio Rao wrote: > > 2010/1/25 Oliver Fromme <olli@lurza.secnetix.de>: > > > [...] > > > > May you revert r202387, 202441 and 202534 and see if it does make a difference? > > > > > > Yes, reverting those fixes the problem. The CPU statistics > > > are back to normal. > > > > With the latest current, may you please provide a verbose dmesg? > > May you also provide a %sysctl kern.timecounter? > > Is a current of last week (2010-01-18) sufficient? I haven't > seen any commits to the clock subsystem in the meantime, so > it should be sufficient. Verbose dmesg, sysctl and much more > can be found here: Thanks a lot for the report. I see this: RTC BIOS diagnostic error 80<clock_battery> which means basically atrtc has been disabled and thus profhz and stathz fellback. The other traces I got are someway similar (I see values of hz which go in this direction). What needs to happen is that, if atrtc is not available apic should be choosen actually. I will try to make a patch very soon. Attilio -- Peace can only be achieved by understanding - A. Einstein
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3bbf2fe11001250858p2516e20y95576eb3cc59764c>