2.6.24 still causes 1000Hz wakeups on x86_64
jonathan.underwood at gmail.com
Sun Mar 16 01:21:25 UTC 2008
On 15/03/2008, Mike Cronenworth <mike at cchtml.com> wrote:
> What's the deal? The tickless kernel shouldn't be causing wakeups at
> idle, correct?
> On a Pentium 4 3ghz HT machine, 2.6.24 shows very little wakeups
> occuring, but on my Core 2 Duo 2.1ghz machine I'm getting over 1000
> wakeups and it's keeping one core fully awake.
> powertop output:
> Top causes for wakeups:
> 88.4% (1000.0) events/0 : run_workqueue (ir_timer)
> 5.5% ( 62.0) <interrupt> : uhci_hcd:usb3, ahci, nvidia
> 1.2% ( 13.2) <interrupt> : uhci_hcd:usb4, libata
> 1.1% ( 12.0) <kernel module> : usb_hcd_poll_rh_status (rh_timer_func)
> 0.9% ( 10.0) <kernel core> : ehci_irq (ehci_watchdog)
> 0.5% ( 5.2) thunderbird-bin : futex_wait (hrtimer_wakeup)
> On my P4 machine, "events/0 : run_workqueue" is further down on the list
> and isn't causing major wakeups.
> Both test machines are using Fedora 8. Latest updates. XFCE desktop. No
> process or applications are running (besides Thunderbird, obviously).
> Google searching resulted in nothing positive. Am I the only one?
On my intel core 2 duo laptop, I don't see the problem you report:
Wakeups-from-idle per second : 177.2 interval: 10.0s
no ACPI power usage estimate available
Top causes for wakeups:
35.2% ( 58.5) npviewer.bin : schedule_timeout (process_timeout)
26.9% ( 44.6) <kernel IPI> : Rescheduling interrupts
7.8% ( 13.0) <interrupt> : iwl3945
5.2% ( 8.6) <interrupt> : libata
3.8% ( 6.3) file-roller : schedule_timeout (process_timeout)
2.8% ( 4.7) X : do_setitimer (it_real_fn)
But you're probably seeing a legitimate bug (perhaps incorrect boot
time selection of clock source or something). It's probably worth
reporting this as a bug with as much detail about your hardware as you
can (lspci -vv etc).
> fedora-list mailing list
> fedora-list at redhat.com
> To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
More information about the users