Date: Sun, 10 May 2009 13:56:01 -0700 From: perryh@pluto.rain.com To: onemda@gmail.com Cc: freebsd-questions@freebsd.org Subject: Re: how to fix "interrupt storm" Message-ID: <4a073f61.RrxzC7F5VdEWczgd%perryh@pluto.rain.com> In-Reply-To: <3a142e750905101247h7f01d6a4obbcbc81d3ae4b7a7@mail.gmail.com> References: <4a04702a.ZaIfHAUzw/YexVK2%perryh@pluto.rain.com> <3a142e750905090746g5324d8ffl1ea10645c0e5f45c@mail.gmail.com> <4a05f963.OI3CMfJ3/j2hbi4D%perryh@pluto.rain.com> <3a142e750905100452s73aa391bjcdc8fea49636ee37@mail.gmail.com> <4a070969.mYZsiV8emLhJVvVg%perryh@pluto.rain.com> <3a142e750905101247h7f01d6a4obbcbc81d3ae4b7a7@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
"Paul B. Mahol" <onemda@gmail.com> wrote: > On 5/10/09, perryh@pluto.rain.com <perryh@pluto.rain.com> wrote: ... > >> >> > interrupt storm detected on "irq9:"; throttling interrupt source > >> >> > interrupt storm detected on "irq9:"; throttling interrupt source > >> >> > ad6: FAILURE - SET_MULTI status=51<READY,DSC,ERROR> > >> >> > error=4<ABORTED> > >> >> > ad6: TIMEOUT - READ_DMA retrying (1 retry left) LBA=136936 > >> >> > > >> >> > etc. etc. until I killed it with ^C ... > >> >> > >> >> Output of "vmstat -i"? > >> > > >> > $ vmstat -i > >> > interrupt total rate > >> > irq0: clk 497386851 1004 > >> > irq1: atkbd0 2491 0 > >> > irq3: xl0 2030 0 > >> > irq6: fdc0 11 0 > >> > irq7: ppbus0 ppc0 1 0 > >> > irq8: rtc 63654324 128 > >> > irq9: uhci0+ 166216 0 > >> > >> uhci0 is doing strange things, what usb device are connected? > > > > There are no USB devices connected. I think those must actually > > be atapci1 interrupts, since irq9 is where dmesg reported it. > > Try editing /boot/device.hints lines with irq or adding similar > lines ... How would I go about figuring out what to add or change? I suppose I want to move either uhci0 or atapci1 to an unused irq, but my recollection is that I don't have unlimited choice in the matter because the IRQ used by a particular PCI device -- or at least the set available for assignment -- is determined by how the motherboard is wired. Granted it's been several years since I was into PCI at this level.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4a073f61.RrxzC7F5VdEWczgd%perryh>