On 03/13/2010 05:08 PM, Ancoron Luciferis wrote:
> On 03/13/2010 04:28 PM, Stefan Bader wrote:
>> I would check whether the number of timer interrupts on your system increments.
>> You can see this with 'watch cat /proc/interrupts" (its IRQ0).
>> If that number does not increase quite often, you might try to boot with
>> "acpi_skip_timer_override" on the grub/kernel commandline.
> Well, actually it does increase normally (compared to a similar karmic
> system which doesn't yield such lags):
> CPU0 CPU1 CPU2 CPU3
> 0: 72 133 5761577 13517 IO-APIC-edge timer
> Well, I just might give it a try and see whether it does anything. On
> the other side I noticed there are some messages at dmesg regarding ACPI:
> [ 0.720325] ACPI: WMI: Mapper loaded
> [ 0.720362] PCI: Using ACPI for IRQ routing
> [ 0.720404] pci 0000:00:00.0: BAR 3: address space collision on of
> device [0xe0000000-0xffffffff]
> [ 0.720445] pci 0000:00:00.0: BAR 3: can't allocate resource
> [ 11.468624] hub 8-0:1.0: USB hub found
> [ 11.468629] hub 8-0:1.0: 4 ports detected
> [ 11.469101] ACPI: I/O resource piix4_smbus [0xb00-0xb07] conflicts
> with ACPI region SOR1 [0xb00-0xb0f]
> [ 11.469115] ACPI: If an ACPI driver is available for this device, you
> should use it instead of the native driver
> [ 11.479951] EDAC amd64: This node reports that Memory ECC is
> currently disabled, set F3x44 (0000:00:18.3).
> [ 11.479959] EDAC amd64: ECC disabled in the BIOS or no ECC
> capability, module will not load.
> [ 11.479960] Either enable ECC checking or force module loading by
> setting 'ecc_enable_override'.
> [ 11.479961] (Note that use of the override may cause unknown side
> [ 11.479973] amd64_edac: probe of 0000:00:18.2 failed with error -22
> Is this something I should be worried of?
Well, I'm currently not experiencing any big lags within the last 36
hours. Top ten of those:
[1268618535.883009998] missed 1016 ms
[1268601187.965986470] missed 959 ms
[1268661073.942935729] missed 958 ms
[1268654844.350597765] missed 919 ms
[1268667242.875538143] missed 908 ms
[1268621185.118586887] missed 898 ms
[1268658402.354881825] missed 895 ms
[1268649717.540822470] missed 895 ms
[1268666098.517713161] missed 893 ms
[1268601088.309843968] missed 892 ms
and the system was almost completely idle.
What I changed was to disable Cool'n'Quiet in the BIOS. I thought the
problems with that were solved already. But it seems they are not.
Does this behavior point someone to anything special? I really would
like to use the CPU scaling features as it will get very hot when the
kernel-team mailing list