2008-02-02 15:05:41

by Guennadi Liakhovetski

[permalink] [raw]
Subject: [X86 ACPI SMP] system slow after cold start if processor module loaded

Hi

Don't know if this should be considered a bug-report, or just another ACPI
bug on my system. In any case, if the CONFIG_ACPI_PROCESSOR is enabled, on
first boot after a power off the system is slow. Not like if only 1 of 2
CPUs actually was running, much slower yet. Respectively, if the module is
built into the kernel, it is slow more or less from the beginning - I
guess, from the moment, when the module is initialized, if built as a
module, it is slow after the module gets loaded. Also funny, it is only
slow on the first boot, after a reboot into the same kernel it runs
normally. The system is a dual P-II@400MHz, Compaq AP400. It is known to
have various ACPI bugs, so, this is just another one of them. No idea
whether or not this shall and can be fixed. At least wanted to document it
in case someone has a similar problem.

Thanks
Guennadi
---
Guennadi Liakhovetski


2008-02-05 23:15:32

by Andrew Morton

[permalink] [raw]
Subject: Re: [X86 ACPI SMP] system slow after cold start if processor module loaded

On Sat, 2 Feb 2008 16:05:41 +0100 (CET)
Guennadi Liakhovetski <[email protected]> wrote:

> Hi
>
> Don't know if this should be considered a bug-report, or just another ACPI
> bug on my system. In any case, if the CONFIG_ACPI_PROCESSOR is enabled, on
> first boot after a power off the system is slow. Not like if only 1 of 2
> CPUs actually was running, much slower yet. Respectively, if the module is
> built into the kernel, it is slow more or less from the beginning - I
> guess, from the moment, when the module is initialized, if built as a
> module, it is slow after the module gets loaded. Also funny, it is only
> slow on the first boot, after a reboot into the same kernel it runs
> normally. The system is a dual P-II@400MHz, Compaq AP400. It is known to
> have various ACPI bugs, so, this is just another one of them. No idea
> whether or not this shall and can be fixed. At least wanted to document it
> in case someone has a similar problem.
>

(added linux-acpi)

If nothing happens with this (likely) then please raise a report against
acpi at bugzilla.kernel.org, thanks.