Bug 1643

Summary:	Fan status not working in D600 laptop
Product:	ACPI	Reporter:	ow (ow.mun.heng)
Component:	Power-Fan	Assignee:	Luming Yu (luming.yu)
Status:	REJECTED INVALID
Severity:	normal	CC:	acpi-bugzilla, ow.mun.heng
Priority:	P2
Hardware:	i386
OS:	Linux
Kernel Version:	2.4.23-cpufreq-patched	Subsystem:
Regression:	---	Bisected commit-id:
Attachments:	output of acpidmp output of acpidmp DSDT \| acpidiasm output of acpidmp FACP \| acpitbl output of biosdecode My kernel config file output of dmesg output of dmidecode Zip of dsdt.aml/hex/dsl output of iasl -tc cat /proc/interuppts output of lspci -vv

Description ow 2003-12-04 01:24:38 UTC

Distribution: Redhat linux 9.0
Hardware Environment: Dell D600 Laptop
Software Environment: 
Problem Description: Cannot get fan status and cannot get the fan to turn off. 
Fan is On _all_ the time. Suspect this is giving me poor battery life. i8k 
utils will not work. i8kfan does nothing. insmod i8k force=1 can get /proc/i8k 
but nothing can change it. 

#ls -laR of /proc/acpi/fan 
Total 0


Steps to reproduce: 
Just cat /proc/acpi/fan


Attachment : I attached a whole lot of stuffs. My DSDT table in hex, dsl, aml, 
output of lspic -vv, dmidecode,biosdecode, acpidmp, config_file (all acpi 
related stuffs are compiled into kernel and not as module. This does not 
include the Toshiba/asus stuffs)

Comment 1 ow 2003-12-04 01:27:15 UTC

Created attachment 1617 [details]
output of acpidmp

Comment 2 ow 2003-12-04 01:28:09 UTC

Created attachment 1618 [details]
output of acpidmp DSDT | acpidiasm

Comment 3 ow 2003-12-04 01:28:56 UTC

Created attachment 1619 [details]
output of acpidmp FACP | acpitbl

Comment 4 ow 2003-12-04 01:29:28 UTC

Created attachment 1620 [details]
output of biosdecode

Comment 5 ow 2003-12-04 01:30:03 UTC

Created attachment 1621 [details]
My kernel config file

Comment 6 ow 2003-12-04 01:30:28 UTC

Created attachment 1622 [details]
output of dmesg

Comment 7 ow 2003-12-04 01:31:14 UTC

Created attachment 1623 [details]
output of dmidecode

Comment 8 ow 2003-12-04 01:32:24 UTC

Created attachment 1624 [details]
Zip of dsdt.aml/hex/dsl

Comment 9 ow 2003-12-04 01:33:00 UTC

Created attachment 1625 [details]
output of iasl -tc

Comment 10 ow 2003-12-04 01:33:31 UTC

Created attachment 1626 [details]
cat  /proc/interuppts

Comment 11 ow 2003-12-04 01:34:08 UTC

Created attachment 1627 [details]
output of lspci -vv

Comment 12 Luming Yu 2003-12-11 18:33:37 UTC

Does this laptop works well on Windows? 

From your DSDT, _STA method of BAT0 looks well.
Below statement will determine whether BAT0/BAT1 is present or not.

    And (MIS0, 0x02, Local0)

And below code could be the only chance to change above decision:
        Device (PCI0)
        {
         ....
            Method (_INI, 0, NotSerialized)
            {
                Store (SMI (0x98, 0x00), MIS0)
                And (MIS0, Or (0x01, Or (0x02, 0x10)), MIS0)
                ...
            }  
          ....
        }

Obviously, SMI(0x98,0x00) is the root.

    Method (SMI, 2, NotSerialized)
    {
        Acquire (SMIX, 0xFFFF)
        Store (Arg1, \_SB.SMIA)
        Store (Arg0, \_SB.SMIC)
        Store (\_SB.SMIC, Local0)
        While (LNot (LEqual (Local0, 0x00)))
        {
            Store (\_SB.SMIC, Local0)
        }

        Store (\_SB.SMIA, Local1)
        Release (SMIX)
        Return (Local1)
    }

Obviously, the writter of above code try to get the correct battery present 
status after the while loop. 

We need to verify \_SB.SMIA has the correct value, before we could determine 
whether this is Linux ACPI bug!

Comment 13 ow 2003-12-14 21:31:23 UTC

This laptop works Ok with Windows (XP Pro) The fan does not turn on all the 
time at all. In fact, the fan can be controlled via a third party util called 
i8kFanGui. It can be turned up (3000+ rpm), turned off (0rpm) and left as is.

I also noticed that in the cat /proc/acpi/thermal*/THM/cooling_mode  gives

<not supported>

I'm not sure if that's anything. Also read about turning acpi=off when loading 
to determine if ACPI is playing a role. Will try that..

Comment 14 ow 2003-12-21 17:13:53 UTC

Booting with ACPI=Off, fan will not be turned on until laptop
temperature is high and BIOS kicks in and turns on fan.

Tested with i8k modules, the i8kctl program is able to control
the fan with much success.

I'm pretty happy with that.

Comment 15 Luming Yu 2004-01-01 21:40:35 UTC

  The DSL file shows that your laptop don't have FAN devices and Active 
cooling mode support. So it's impossbile to let ACPI code to control your FAN.
So it is not linux ACPI kernel bug.  Maybe windows is using SMM BIOS code to 
support FAN rather than ACPI code too.

  Regarding cooling_mode gives <not supported>, it just give false information.
I have a patch filed at bug 1770

  So, I will close this bug.

Comment 16 Rafael J. Wysocki 2008-10-05 03:49:06 UTC

On Sunday, 5 of October 2008, Takashi Iwai wrote:
> At Sat,  4 Oct 2008 19:32:55 +0200 (CEST),
> Rafael J. Wysocki wrote:
> > 
> > Bug-Entry   : http://bugzilla.kernel.org/show_bug.cgi?id=11643
> > Subject             : ALSA sound/core/pcm_native.c:1947: BUG? (err &gt;= 0)
> > Submitter   : sangu <sangu.gnome@gmail.com>
> > Date                : 2008-09-24 16:51 (11 days old)
> 
> This is unlikely a 2.6.27 regression.
> It's a bug to be fixed, though.