Bug 8877

Summary:	sighting of broken acpi bios trip points
Product:	ACPI	Reporter:	Bryon Gill (thepiratekings)
Component:	Power-Thermal	Assignee:	Len Brown (lenb)
Status:	REJECTED INSUFFICIENT_DATA
Severity:	normal	CC:	acpi-bugzilla, lenb, lsmaneuf
Priority:	P1
Hardware:	All
OS:	Linux
Kernel Version:	2.6.22.1-32.fc6	Subsystem:
Regression:	---	Bisected commit-id:

Description Bryon Gill 2007-08-10 18:30:01 UTC

Most recent kernel where this bug did not occur:
Distribution:
Hardware Environment:
Software Environment:
Problem Description:

Steps to reproduce:

Long story short, I need to adjust my trip points to make my machine usable.  The passive cooling kicks in at 50 slowing my computer to a crawl and it shuts off at 60 - during the summer with A/C on it still runs 55-60 most of the time.  I'd give you an acpidump but I don't have it on my machine and I'm afraid getting it and compiling it might actually overheat my computer and cause a shutdown.


[root@brybox THRM]# cat /proc/acpi/thermal_zone/*/*
0 - Active; 1 - Passive
<polling disabled>
state:                   passive 
temperature:             57 C
critical (S5):           60 C
passive:                 50 C: tc1=4 tc2=3 tsp=60 devices=CPU0 
active[0]:               50 C: devices= FAN 
[root@brybox THRM]# 


Just for reference, my computer has been running just fine for over a year with this command in my rc.local:
echo 75:0:64:64:85 > /proc/acpi/thermal_zone/THRM/trip_points


I'll see if I can get an acpidump and update this but count my machine as one affected by the locking of trip point values.  Let me add that the process for changing the damned things is so arcane that you're unlikely to change them without a lot of investigation and learning in the first place.

Comment 1 Len Brown 2007-08-10 22:15:45 UTC

Based on the experience in bug 8842, I expect that this system
has broken thermal events, and its bogus trip-points are being
noticed by Linux because Linux is erroneously polling.

Please boot with acpi=off and collect the acpidump and dmidecode output
to attach to this bug report.

Note that you can also boot with ACPI enabled but with thermal control off
by building with CONFIG_ACPI_THERMAL=n, or if it is a module,
# rmmod thermal

Please boot with ACPI enabled and
# cat /proc/acpi/thermal_zone/*/polling_frequency
and if the distro has enabled polling on you, then:
# echo 0 > /proc/acpi/thermal_zone/*/polling

then please report how the system behaves.

In particular, kill acpid and cat /proc/acpi/event
and monitor /proc/interrupts for the acpi line
to see if there are any ACPI thermal events reported
as the temperature crosses 50 and 60.

Comment 2 Bryon Gill 2007-08-11 15:27:24 UTC

When I boot with acpi=off my network driver breaks.  Where can I get acpidumpand dmidecode?  I'm using a fedora core 6 kernel and the thermal appears to be built in (lsmod | grep thermal is empty).  Polling reports itself as being disabled.

I do get a message telling me that it can't turn the fan on when I boot with the new kernel, it spams the console every couple seconds.  

Hope that's helpful, I'll see if I can get the other information for you later on.

Comment 3 Bryon Gill 2007-08-11 15:28:50 UTC

One more thing- despite what the acpi messages tell me, the fans seem to be running.

Comment 4 Len Brown 2007-08-11 20:13:03 UTC

It is true that fedora uses CONFIG_ACPI_THERMAL=y
and thus you'll not be able to unload that module
when running a fedora kernel.

To best help diagnose and debug a problem here, you'll
need to be able to run a kernel.org kernel --
not just a fedora kernel.

dmidecode can be had from the fedora distribution,
try applications/"Add/Remove Software"
and search for dmidecode

failing that, you can get it from 
www.nongnu.org/dmidecode/ 
which is the 1st hit you'll find if you google it.

apparently acpidump doesn't come with fedora,
you can get it from the latest pmtools here:
http://ftp.kernel.org/pub/linux/kernel/people/lenb/acpi/utils/

re: messages about the fan
please include the complete output dmesg -s64000

Comment 5 Fu Michael 2007-10-18 01:35:52 UTC

ping for response from bug reporter...

Comment 6 Fu Michael 2007-12-03 00:52:02 UTC

reject the bug due to no response from bug reporter. please reopen with the info requested in comment# 4.