Bug 11877 - Critical Trip Point shutdown - HP Pavillion DV8000t
Summary: Critical Trip Point shutdown - HP Pavillion DV8000t
Status: REJECTED INSUFFICIENT_DATA
Alias: None
Product: ACPI
Classification: Unclassified
Component: Power-Thermal (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: ykzhao
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-10-28 10:38 UTC by salatiel.filho
Modified: 2009-02-09 17:20 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.27
Subsystem:
Regression: Yes
Bisected commit-id:


Attachments
acpidump (125.81 KB, application/octet-stream)
2008-11-08 10:49 UTC, salatiel.filho
Details
acpidump --addr 0x0xBFE93E4C --length 0x0100 -o gnvs (125.81 KB, application/octet-stream)
2008-11-09 10:54 UTC, salatiel.filho
Details
gnvs_2.6.25.7 (256 bytes, text/plain)
2008-11-16 06:54 UTC, salatiel.filho
Details

Description salatiel.filho 2008-10-28 10:38:06 UTC
Latest working kernel version: 2.6.21.7
Earliest failing kernel version:
Distribution: Gentoo 
Hardware Environment: HP Pavillion DV8000t
Software Environment:
Problem Description: I have upgraded my system from 2.6.21.7 to 2.6.27 and now my laptop keeps shutting down with ACPI – Critical Trip Point when i try to use any cpu intensive app. i dont think this is a real overheating problem  cause i made a test in windows running stress programs for a few hours and it  worked just fine.

What else information should i provide to help fix this ?
Comment 1 ykzhao 2008-10-28 18:30:44 UTC
Will you please compile the drivers/acpi/thermal as built-in kernel and add the boot option of "thermal.nocrt=1"?
   After the system is booted, please cat the output of /proc/acpi/thermal/*/*
   

Of course please attach the output of acpidump.
Thanks.
Comment 2 salatiel.filho 2008-11-08 10:49:47 UTC
Created attachment 18740 [details]
acpidump

acpidump
Comment 3 salatiel.filho 2008-11-08 10:50:39 UTC
There is no /proc/acpi/thermal/*/ , so i think the path is: 

root@gentoo ~ # cat /proc/acpi/thermal_zone/TZ0*/*
<setting not supported>
<polling disabled>
state:                   ok
temperature:             55 C
critical (S5):           99 C <disabled>
<setting not supported>
<polling disabled>
state:                   ok
temperature:             27 C
critical (S5):           105 C <disabled>
Comment 4 ykzhao 2008-11-09 05:33:05 UTC
Will you please attach the following output ?
   ./acpidump --addr 0x0xBFE93E4C --length 0x0100 -o gnvs

Please add the boot option of "thermal.nocrt=1". Please try to
use any cpu intensive app and then attach the output of "cat /proc/acpi/thermal_zone/TZ*/*"
Thanks
Comment 5 salatiel.filho 2008-11-09 10:52:48 UTC
Coretemp module shows CPU core0 at 86C and core1 at 82C when i cat this:

<setting not supported>
<polling disabled>
state:                   critical 
temperature:             100 C
critical (S5):           99 C <disabled>
<setting not supported>
<polling disabled>
state:                   ok
temperature:             27 C
critical (S5):           105 C <disabled>
Comment 6 salatiel.filho 2008-11-09 10:54:09 UTC
Created attachment 18753 [details]
acpidump --addr 0x0xBFE93E4C --length 0x0100 -o gnvs 

acpidump --addr 0x0xBFE93E4C --length 0x0100 -o gnvs
Comment 7 Zhang Rui 2008-11-09 19:59:06 UTC
(In reply to comment #6)
> Created an attachment (id=18753) [details]
> acpidump --addr 0x0xBFE93E4C --length 0x0100 -o gnvs 
> 
sorry, typo.

please do the test like this:
0. boot into 2.6.27 kernel
1. "acpidump --addr 0xBFE93E4C --length 0x0100 -o gnvs"
2. "cat /proc/acpi/thermal_zone/*/* > temp"
3. re-do this test in 2.6.21 kernel
then attach the four test results here. :)
Comment 8 Len Brown 2008-11-11 23:05:19 UTC
> coretemp

does this issue go away if the kernel is built with CONFIG_HWMON=n?
Comment 9 salatiel.filho 2008-11-16 06:52:53 UTC
Sorry , since i deleted 2.6.21 kernel , i tested on 2.6.25.7 cause i already had the source.
On Kernel 2.6.25.7 , machine does not shutdown . I will try to redownload 2.6.21 but i am o dial up connection so it will take a while.

<setting not supported>
<polling disabled>
state:                   ok
temperature:             89 C
critical (S5):           99 C <disabled>
<setting not supported>
<polling disabled>
state:                   ok
temperature:             27 C
critical (S5):           105 C <disabled>
Comment 10 salatiel.filho 2008-11-16 06:54:07 UTC
Created attachment 18880 [details]
gnvs_2.6.25.7

2.6.25.7 acpidump --addr 0xBFE93E4C --length 0x0100 -o gnvs
Comment 11 salatiel.filho 2008-11-16 06:56:18 UTC
Len , i will do the test  right now.
sorry not post earlier but i could not reboot the machine.

I am having lots of other problems with kernel 2.6.27, and most annoying after the auto reboot is my keyboard/mouse just stop working ...btw , where should i report this ?
Comment 12 ykzhao 2008-11-16 18:34:07 UTC
From the acpidump it seems that the temperature of TZ01 thermal zone is obtained by the evaluating the following ACPI object:
   > Method (_TMP, 0, Serialized)
            {
                If (LEqual (\_SB.PCI0.LPCB.EC0.ECRY, One))
                {
                    If (DTSE)
                    {
                        Store (DTS2, Local1)
                        If (LGreaterEqual (DTS1, DTS2))
                        {
                            Store (DTS1, Local1)
                        }

                        Multiply (Local1, 0x0A, Local0)
                        Add (Local0, 0x0AAC, Local0)
                        Return (Local0)
                    }
                }

                Return (0x0BB8)
            }
    It is related with the DTSE, DTS1,DTS2 object defined in GNVS memory region.
    
    From the gnvs it seems that the initial temperature of TZ01 thermal zone is 55. But why is the 100 degree returned by the TZ01 thermal zone after use the cpu intensive application? It is very strange.
    Will you please do the test as suggested by Len in comment #8?
    thanks.
   
Comment 13 salatiel.filho 2008-11-19 09:53:36 UTC
CONFIG_HWMON=n changed nothing :(, it keeps shutting down.
any other info i can provide ?
Comment 14 ykzhao 2008-11-19 17:35:55 UTC
Thanks for the confirmation. It seems that the problem still exists even when the hardware monitor is disabled.

From the acpidump it seems that the temperature of TZ01 thermal zone is related with the DTSE/DTS1/DTS2. And the DTS1/DTS2 are not changed by OS. It seems that they are changed by BIOS. 
From the problem description it seems that the box will be shutdown after running the CPU intensive application. And we don't know what happens when running the CPU intensive application.Very sorry that this bug can't be fixed by Linux ACPI.

Will you please confirm whether the box will be shutdown if you don't running the CPU intensive application?
Comment 15 salatiel.filho 2008-11-19 17:40:33 UTC
It does not shut down when idle.
it is strange , why does this only happen with new kernels ? Something must be changed...
Anyway , thanks. I will workaround by limiting the cpu speed ... at least for a while.
Comment 16 ykzhao 2008-11-19 19:25:00 UTC
   From the acpidump it seems that this issue is related with the BIOS.
As this issue only happens on the new kernel, will you please use git-bisect to identify which commit the regression is caused by?
   Thanks.
   
Comment 17 Zhang Rui 2008-11-24 22:37:03 UTC
ping salatiel.
Comment 18 salatiel.filho 2008-11-24 23:40:25 UTC
pong. I was sure i had answer this thread :)
I have no idea how to use git bisector :(
Comment 19 Zhang Rui 2008-11-25 00:11:42 UTC
http://www.lesswatts.org/projects/acpi/debug.php
Debug: How to Isolate Linux ACPI Issues

there are several links about how to use git-bisect. :)

maybe this would help. :)
Comment 20 Zhang Rui 2008-12-08 00:23:10 UTC
salatiel, does the info in comment #19 help?
Comment 21 salatiel.filho 2008-12-08 04:38:40 UTC
Hi Zhang , i am in a trip right now, i will post some info as soon as i get back.
Comment 22 Zhang Rui 2008-12-28 17:52:23 UTC
salatiel, are you back and ready for the git-bisect? :)
Comment 23 salatiel.filho 2008-12-30 06:44:41 UTC
Hi , i am back now , sorry , but i almost forgot this. Since i  installed ubuntu 8.10 last week i didn`t have one single "trip point" shutdown. If it is still need i can start bisect next monday , after the parties :)
Comment 24 Zhang Rui 2009-01-04 00:49:28 UTC
Hi, salatiel,
can you reproduce this bug any more?
I'll close this bug if it's not reproducible.
Comment 25 ykzhao 2009-02-09 17:20:34 UTC
Ping Salatiel....
    As there is no response for more than one month, the bug will be rejected.
    If the problem still exists, please use the git-bisect to identify the issue and attach the output of bisect. 

Note You need to log in before you can comment on or make changes to this bug.