Re: High interrupt when using 'dd' with Xen DOM0 kernel

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Sat, May 02, 2009 at 10:15:16PM +0200, Jean-Edouard Babin wrote:
> >> Hello,
> >>
> >> I upgraded from a quite old (november) NetBSD 4.99 i386 to NetBSD 5
> >> amd64. (with latest XEN33 packages)
> >
> > were you using xen33 or xen31 before ?
> 
> It was 3.3.0
> 
> >> I tried dd if=/dev/zero of=DomU.img bs=1m count=10240 to make a new
> >> image to create a domU
> >> but this make interrupt to go to 97%
> >> CPU states:  1.0% user,  0.0% nice,  0.0% system, 97.1% interrupt,  2.0% 
> >> idle
> >> 0 root    126    0    0K  12M pgdaemon  5:36 76.07% 76.07% [system]
> >
> > please try
> > systat vm
> > to see which interrupt rate is high ?
> 
> it seems that the highest int. comes from  ioapic0 pin 20
> 
> # dmesg | grep ioapi | grep 20
> ioapic0 at mainbus0 apid 2: pa 0xfec00000, virtual wire mode, version
> 20, 24 pins
> ioapic0: pin 20 attached to pci3 device 0 INT_A (type 0<type=0> flags
> f<pol=3=Act Lo,trig=3=Level>)
> ioapic0: pin 20 attached to pci3 device 1 INT_D (type 0<type=0> flags
> f<pol=3=Act Lo,trig=3=Level>)
> ioapic0: int20 1a9d8<vector=d8,delmode=1,logical,actlo,level,masked,dest=0>
> 1000000<target=1>
> satalink0: using ioapic0 pin 20, event channel 9 for native-PCI interrupt
> 
> satalink0 at pci3 dev 0 function 0
> tlp0 at pci3 dev 1 function 0: Lite-On 82C169 Ethernet, pass 2.0
> 
> I guess that it's INT_D and not INT_A (I don't know if there is a way
> to see that as sysstat only report pin number).
> 
> Only wd0 is connected to satalink0, so I tried on another harddrive
> and did not get the same problem.

is it another harddrive connected to the same SATA port of the
controller, or another harddrive on another port (or another controller) ?

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--