high system load when using i915kms
Kevin Oberman
rkoberman at gmail.com
Fri Mar 8 19:38:14 UTC 2013
On Fri, Mar 8, 2013 at 10:03 AM, Max Brazhnikov <makc at freebsd.org> wrote:
> On Thu, 07 Mar 2013 16:47:23 +0200 Konstantin Belousov wrote:
> > On Wed, Mar 06, 2013 at 10:56:52AM +0000, Max Brazhnikov wrote:
> > > On Wed, 6 Mar 2013 08:15:37 +0200 Konstantin Belousov wrote:
> > > > On Tue, Mar 05, 2013 at 12:24:51PM +0000, Max Brazhnikov wrote:
> > > > > Hi,
> > > > >
> > > > > I've switched recently to new xorg and kms driver. Sometimes after
> booting
> > > my
> > > > > system (9.1-STABLE #0 r245741 amd64) shows high load:
> > > > >
> > > > > interrupt total rate
> > > > > irq1: atkbd0 612 0
> > > > > irq9: acpi0 3693 1
> > > > > irq12: psm0 7512 2
> > > > > irq16: uhci0 uhci3 377172955 144069
> > > > > irq20: hpet0 2923357 1116
> > > > > irq23: uhci1 ehci1 47432 18
> > > > > irq256: hdac0 80799 30
> > > > > irq257: alc0 78474 29
> > > > > irq258: iwn0 19994 7
> > > > > irq259: ahci0 100016 38
> > > > > irq260: vgapci0 31250 11
> > > > > Total 380466094 145327
> > > > >
> > > > > I've never seen this with old xorg. The problem is not always
> > > reproducible,
> > > > > but it's enough just to load i915kms module without staring X to
> trigger
> > > it.
> > > > >
> > > > > Any idea?
> > > >
> > > > So what is the complain ? Do you meaning that loading i915kms causes
> > > > the spike in the interrupt rate on the irq16 line ?
> > >
> > > I suspect it since the only change was update to newer xorg.
> > Do you mean that the same kernel was kept,
> > and only usermode components upgraded ?
>
> Correct. The kernel and world were built from r245741 more than month ago.
> I used prebuilt Xorg packages from latest experimental ports, so the
> changes
> in the system are minimal.
>
> > This is plain impossible to cause the effect you described.
> > >
> > > > What is the graphics part and the south bridge you are using ? Show
> > > > the pciconf -lvc output.
> > >
> > > http://people.freebsd.org/~makc/pciconf.output
> >
> > I have exactly the same GM45 chipset in my laptop.
> >
> > BTW, is the vmstat output you demonstrated in the first message, was for
> > the system with running X ? I am asking about the presence of the
> rendering
> > activity on the display, which would explain the significant count of
> > the interrupts from GPU.
>
> Yes, it was under KDE session, however the problem can be observe with
> plain X
> and xterm or even without them:
> # cat test.sh
> #!/bin/sh
> vmstat -i
> kldload i915kms
> sleep 10
> vmstat -i
> xinit &
> sleep 10
> vmstat -i
>
> interrupt total rate
> irq1: atkbd0 169 3
> irq9: acpi0 72 1
> irq12: psm0 24 0
> irq16: uhci0 uhci3 25 0
> irq19: ehci0 uhci2 2 0
> irq20: hpet0 20351 407
> irq23: uhci1 ehci1 95 1
> irq256: hdac0 71 1
> irq257: alc0 347 6
> irq258: iwn0 1624 32
> irq259: ahci0 4095 81
> Total 26875 537
> interrupt total rate
> irq1: atkbd0 169 2
> irq9: acpi0 72 1
> irq12: psm0 24 0
> irq16: uhci0 uhci3 1533470 24733
> irq19: ehci0 uhci2 2 0
> irq20: hpet0 33165 534
> irq23: uhci1 ehci1 95 1
> irq256: hdac0 71 1
> irq257: alc0 509 8
> irq258: iwn0 2090 33
> irq259: ahci0 4273 68
> irq260: vgapci0 1 0
> Total 1573941 25386
> interrupt total rate
> irq1: atkbd0 187 2
> irq9: acpi0 94 1
> irq12: psm0 24 0
> irq16: uhci0 uhci3 3056916 42457
> irq19: ehci0 uhci2 2 0
> irq20: hpet0 44423 616
> irq23: uhci1 ehci1 95 1
> irq256: hdac0 71 0
> irq257: alc0 634 8
> irq258: iwn0 2494 34
> irq259: ahci0 4772 66
> irq260: vgapci0 10 0
> Total 3109722 43190
>
> PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU
> COMMAND
> 11 root 2 155 ki31 0K 32K RUN 1 4:16 120.35% idle
> 12 root 18 -84 - 0K 288K WAIT 0 0:57 76.34% intr
>
> I've got this after second boot today, although I couldn't reproduce it
> yesterday even after ten attempts. But sometimes it's quite nasty and I
> have
> to reboot the system several times to get rid of it.
>
> Max
>
So the issue is that that the interrupts from one or another of the USB
devices has exploded from near zero to around 40K when the kernel module is
loaded?
A couple of possibly irrelevant questions. Do you normally manually load
the module? I did not research the issue, but when I manually load the
module I was seeing things just grind to a halt. If I started Gnome, the
module was loaded automatically by X, and things worked.
Why loading the Intel KMS module would cause a massive increase in
interrupts on a USB interface completely baffles me, but I suspect some
sort of race is going on when the module is pre-loaded.
--
R. Kevin Oberman, Network Engineer
E-mail: rkoberman at gmail.com
More information about the freebsd-x11
mailing list