Bug 12910

Summary:	ACPI limits CPU to 800Mhz
Product:	ACPI	Reporter:	Rafael J. Wysocki (rjw)
Component:	Power-Thermal	Assignee:	Len Brown (lenb)
Status:	CLOSED INVALID
Severity:	normal	CC:	acpi-bugzilla, dmitry.torokhov, rui.zhang
Priority:	P1
Hardware:	All
OS:	Linux
Kernel Version:	2.6.29-rc	Subsystem:
Regression:	Yes	Bisected commit-id:
Bug Depends on:
Bug Blocks:	12398

Description Rafael J. Wysocki 2009-03-21 11:13:25 UTC

Subject    : 2.6.29-rcX: wierd iteractions with CPUFREQ
Submitter  : Dmitry Torokhov <dmitry.torokhov@gmail.com>
Date       : 2009-03-19 5:57
References : http://marc.info/?l=linux-kernel&m=123744265002098&w=4

This entry is being used for tracking a regression from 2.6.28.  Please don't
close it until the problem is fixed in the mainline.

Comment 1 Alan 2009-03-21 12:44:09 UTC

Can you confirm going back to older kernels goes back to the old behaviour ? If so which rc did it break ?

Comment 2 ykzhao 2009-03-22 07:01:04 UTC

Hi, Dmitry
    Will you please attach the output of acpidump, dmesg?
    It will be great if you can verify what Alan said in comment #1.
    Thanks.

Comment 3 Zhang Rui 2009-03-23 00:58:18 UTC

please verify if this is a duplicate of bug #12830

Comment 4 Zhang Rui 2009-03-23 01:01:13 UTC

sorry, it's not.
on this laptop, the problem is that no ACPI events when overheating, which results in a low cpu frequency.

Comment 5 Len Brown 2009-03-23 18:45:34 UTC

> acpi_processor_ppc_has_changed

this is the BIOS instructing us to run slower,
so AFAICS we are doing the right thing.

We've seen several machines with BIOS that use PPC update
to manage thermals, eg the T60.  What box is this, and
where is its acpidump?  Did it not heat up in previous releases?
Is the fan spinning at full speed when this happens?
(has it recently become clogged?)

Comment 6 Zhang Rui 2009-03-30 05:50:43 UTC

ping Dmitry.

Comment 7 Dmitry Torokhov 2009-03-30 09:04:32 UTC

Ignore me, please. It was indeed _PPC silently limiting the frequency, apparently because nvidia chip is overheating. Nothing to do with 2.6.29 except that I was trying to go off binary module and back to free nv driver at the same time and apparently nv keeps the chip at max performance at all times.