Date: Sun, 14 Oct 2007 23:14:52 +0200 From: Erik Trulsson <ertr1013@student.uu.se> To: Abdullah Ibn Hamad Al-Marri <wearabnet@yahoo.ca> Cc: Peter Jeremy <peterjeremy@optushome.com.au>, freebsd-acpi@freebsd.org, freebsd-stable@freebsd.org Subject: Re: FreeBSD 7.0 interrupt storm with irq0: clk Message-ID: <20071014211452.GA72643@owl.midgard.homeip.net> In-Reply-To: <58790.8375.qm@web33704.mail.mud.yahoo.com> References: <58790.8375.qm@web33704.mail.mud.yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, Oct 14, 2007 at 01:03:44PM -0700, Abdullah Ibn Hamad Al-Marri wrote: > On 2007-Oct-14 15:36:39 +0300, Abdullah Ibn Hamad Al-Marri > <almarrie@gmail.com> wrote: > >On 10/13/07, Nate Lawson <nate@root.org> wrote: > >> Rolf Witt wrote: > >> > Abdullah Ibn Hamad Al-Marri schrieb: > >> >> Hello, > >> >> > >> >> I'm getting interrupt storm lately. > ... > >But still same issue with irq0: clk but less now. > > > >IM# vmstat -i > >interrupt total rate > >irq0: clk 37009569 1000 > > >So far, you haven't provided any evidence of any "interrupt storm" or > >how it might be related to acpi. And 7.0 isn't stable yet. > > >The default HZ is 1000 so unless you explicitly change it in your > >kernel config, you should have 1000 clock interrupts/sec. > > >-- > >Peter Jeremy > > > > > Actually no more interrupt after I added these options back to my kernel while it's UP not SMP, and vmstat output has changed totally. > > # To make an SMP kernel, the next two lines are needed > options SMP # Symmetric MultiProcessor Kernel > device apic # I/O APIC > > IM# vmstat -i > interrupt total rate > irq14: ata0 4837 0 > irq23: fxp0 2203298 86 > cpu0: timer 50755929 1999 > Total 52964064 2086 > > I even don't see irq0 in vmstat at all. > > Any hints? It looks like different timecounters are used depending on your kernel options. In the first case you probably used TSC or i8254 as timecounter while in the second (where you get timer interrupts on cpu0 instead of on irq0) you use ACPI as timecounter. The output of 'sysctl kern.timecounter.choice' for both cases might be illuminating. In any case there is still no indication of any interrupt storm. -- <Insert your favourite quote here.> Erik Trulsson ertr1013@student.uu.se
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20071014211452.GA72643>