[ltp] HPET timer on thinkpads?

Dan Sawyer linux-thinkpad@linux-thinkpad.org
Fri, 26 Oct 2007 06:09:45 -0700 

This is a multi-part message in MIME format.
--------------020500080801060302050801
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Leon,

Thank is an interesting number. My T41 uses about 500 mw with the 
radeonfb and 2200 without. Try something else, do you have the program 
'glxgears'. Run that as root. With the radeon driver I get about 
1100/sec. With radeonfb it is about 300. I seem to have a choice, use of 
the acceleration H/W or low power suspend. - Dan

Leon Brokken wrote:
> Thanks Dan, here's the output of the script to check the power drain
> during suspend:
>
> leon@shpritsz:~/myscripts$ cat /var/log/battery.log
> Fri Oct 26 10:27:35 CEST 2007
> before: 29500 mWh
> after: 28380 mWh
> diff: -1120 mWh
> seconds: 3250 sec
> result: -1240 mW
> Your model seems to be affected.
>
> I have added T41 2373-8RG to the 'affected' list on thinkwiki.
> I will check later if "radeon_force_sleep=1" boot parameter makes a
> difference.
>
> Cheers, Leon.
>
>
> Dan Sawyer wrote:
>   
>> Leon,
>>
>> Thank you very much. That was very helpful.
>>
>> There are a couple of ways you can check. There is a script on the
>> think pad wike to test power.
>>
>> The other is to read the battery charge (cat
>> /proc/acpi/battery/BAT0/*) suspend the machine for 1 or more hours and
>> then reread the charge. The difference should be about 500mw per hour.
>>
>> Dan
>>
>> Leon Brokken wrote:
>>     
>>> Hi Dan,
>>>
>>> uhm, how would I check that, without waiting for the batt to run out?
>>>  .config is attached.
>>>
>>> Cheers, Leon.
>>>
>>>
>>> Dan Sawyer wrote:
>>>   
>>>       
>>>> Leon,
>>>>
>>>> Does the low power suspend work on your configuration? If so can you
>>>> your .config script.
>>>>
>>>> Thanks - Dan
>>>>
>>>> Leon Brokken wrote:
>>>>     
>>>>         
>>>>> Hi Shem,
>>>>>
>>>>> Shem Multinymous wrote:
>>>>>   
>>>>>       
>>>>>           
>>>>>> Hi Leon,
>>>>>>
>>>>>> On 10/23/07, Leon Brokken <leon.brokken@med.lu.se> wrote:
>>>>>>   
>>>>>>     
>>>>>>         
>>>>>>             
>>>>>>> Could that explain the hangs every now and then, when my T41 doesn't
>>>>>>> accept any input any longer? Processes keep running but there is no info
>>>>>>> from fan, batt or temp available (judged from conky system monitor).
>>>>>>> Even the power button doesn't work anymore, I need to power off by
>>>>>>> unplugging AC and removing the battery.
>>>>>>>     
>>>>>>>       
>>>>>>>           
>>>>>>>               
>>>>>> Judging by the symptoms, what you're seeing is a hang of the embedded
>>>>>> controller (which is running its own firmware independently of the CPU
>>>>>> and handles all the things you mentioned). This is very dangerous,
>>>>>> since the EC is also in charge of fan control, battery charging
>>>>>> control and emergency thermal shutdown.
>>>>>>
>>>>>> It's important to find out what causes the EC hangs. How often do you
>>>>>> see them? Are they correlated with some activity or configuration?
>>>>>> Does unloading some modules (in particular acpi, hdaps, tp_smapi)
>>>>>> prevent them? Are you using an old version of tp_smapi?
>>>>>>   
>>>>>>     
>>>>>>         
>>>>>>             
>>>>> In the meantime I found out that tp_smapi (v0.32) was not loaded. I have
>>>>> now added it to rc.modules so it gets loaded at boot and haven't noticed
>>>>> a problem yet. Before it happpened maybe once every few days, without
>>>>> any obvious pattern. I have a T41, Slackware-current, kernel 2.6.23.1. I
>>>>> never noticed similar issues with 2.6.22.
>>>>>
>>>>> Anyway, here's some output from dmesg:
>>>>>
>>>>> leon@shpritsz:~$ dmesg|grep -i thinkpad_ec
>>>>> thinkpad_ec: thinkpad_ec 0.32 loaded.
>>>>>
>>>>> leon@shpritsz:~$ dmesg|grep -i smapi
>>>>> tp_smapi 0.32 loading...
>>>>> tp_smapi successfully loaded (smapi_port=0xb2).
>>>>>
>>>>> leon@shpritsz:~$ dmesg|grep -i hdaps
>>>>> hdaps: initial mode latch is 0x01
>>>>> hdaps: setting ec_rate=250, filter_order=2
>>>>> hdaps: fake_data_mode set to 0
>>>>> hdaps: device successfully initialized.
>>>>> input: ThinkPad HDAPS joystick emulation as
>>>>> /devices/platform/hdaps/input/input9
>>>>> input: ThinkPad HDAPS accelerometer data as
>>>>> /devices/platform/hdaps/input/input10
>>>>> hdaps: driver successfully loaded.
>>>>> leon@shpritsz:~$
>>>>>
>>>>> leon@shpritsz:~$ dmesg|grep thinkpad_acpi
>>>>> thinkpad_acpi: ThinkPad ACPI Extras v0.18-20071013
>>>>> thinkpad_acpi: http://ibm-acpi.sf.net/
>>>>> thinkpad_acpi: ThinkPad BIOS 1RETDRWW (3.23 ), EC 1RHT71WW-3.04
>>>>> thinkpad_acpi: IBM ThinkPad T41
>>>>> leon@shpritsz:~$
>>>>>
>>>>> I'll report back if it happens again.
>>>>>
>>>>> Thanks, Leon.
>>>>>
>>>>>   
>>>>>       
>>>>>           
>>>   
>>>       
>
>   

--------------020500080801060302050801
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Leon,<br>
<br>
Thank is an interesting number. My T41 uses about 500 mw with the
radeonfb and 2200 without. Try something else, do you have the program
'glxgears'. Run that as root. With the radeon driver I get about
1100/sec. With radeonfb it is about 300. I seem to have a choice, use
of the acceleration H/W or low power suspend. - Dan <br>
<br>
Leon Brokken wrote:
<blockquote cite="mid:4721B681.8040204@med.lu.se" type="cite">
  <pre wrap="">Thanks Dan, here's the output of the script to check the power drain
during suspend:

leon@shpritsz:~/myscripts$ cat /var/log/battery.log
Fri Oct 26 10:27:35 CEST 2007
before: 29500 mWh
after: 28380 mWh
diff: -1120 mWh
seconds: 3250 sec
result: -1240 mW
Your model seems to be affected.

I have added T41 2373-8RG to the 'affected' list on thinkwiki.
I will check later if "radeon_force_sleep=1" boot parameter makes a
difference.

Cheers, Leon.


Dan Sawyer wrote:
  </pre>
  <blockquote type="cite">
    <pre wrap="">Leon,

Thank you very much. That was very helpful.

There are a couple of ways you can check. There is a script on the
think pad wike to test power.

The other is to read the battery charge (cat
/proc/acpi/battery/BAT0/*) suspend the machine for 1 or more hours and
then reread the charge. The difference should be about 500mw per hour.

Dan

Leon Brokken wrote:
    </pre>
    <blockquote type="cite">
      <pre wrap="">Hi Dan,

uhm, how would I check that, without waiting for the batt to run out?
 .config is attached.

Cheers, Leon.


Dan Sawyer wrote:
  
      </pre>
      <blockquote type="cite">
        <pre wrap="">Leon,

Does the low power suspend work on your configuration? If so can you
your .config script.

Thanks - Dan

Leon Brokken wrote:
    
        </pre>
        <blockquote type="cite">
          <pre wrap="">Hi Shem,

Shem Multinymous wrote:
  
      
          </pre>
          <blockquote type="cite">
            <pre wrap="">Hi Leon,

On 10/23/07, Leon Brokken <a class="moz-txt-link-rfc2396E" href="mailto:leon.brokken@med.lu.se">&lt;leon.brokken@med.lu.se&gt;</a> wrote:
  
    
        
            </pre>
            <blockquote type="cite">
              <pre wrap="">Could that explain the hangs every now and then, when my T41 doesn't
accept any input any longer? Processes keep running but there is no info
from fan, batt or temp available (judged from conky system monitor).
Even the power button doesn't work anymore, I need to power off by
unplugging AC and removing the battery.
    
      
          
              </pre>
            </blockquote>
            <pre wrap="">Judging by the symptoms, what you're seeing is a hang of the embedded
controller (which is running its own firmware independently of the CPU
and handles all the things you mentioned). This is very dangerous,
since the EC is also in charge of fan control, battery charging
control and emergency thermal shutdown.

It's important to find out what causes the EC hangs. How often do you
see them? Are they correlated with some activity or configuration?
Does unloading some modules (in particular acpi, hdaps, tp_smapi)
prevent them? Are you using an old version of tp_smapi?
  
    
        
            </pre>
          </blockquote>
          <pre wrap="">In the meantime I found out that tp_smapi (v0.32) was not loaded. I have
now added it to rc.modules so it gets loaded at boot and haven't noticed
a problem yet. Before it happpened maybe once every few days, without
any obvious pattern. I have a T41, Slackware-current, kernel 2.6.23.1. I
never noticed similar issues with 2.6.22.

Anyway, here's some output from dmesg:

leon@shpritsz:~$ dmesg|grep -i thinkpad_ec
thinkpad_ec: thinkpad_ec 0.32 loaded.

leon@shpritsz:~$ dmesg|grep -i smapi
tp_smapi 0.32 loading...
tp_smapi successfully loaded (smapi_port=0xb2).

leon@shpritsz:~$ dmesg|grep -i hdaps
hdaps: initial mode latch is 0x01
hdaps: setting ec_rate=250, filter_order=2
hdaps: fake_data_mode set to 0
hdaps: device successfully initialized.
input: ThinkPad HDAPS joystick emulation as
/devices/platform/hdaps/input/input9
input: ThinkPad HDAPS accelerometer data as
/devices/platform/hdaps/input/input10
hdaps: driver successfully loaded.
leon@shpritsz:~$

leon@shpritsz:~$ dmesg|grep thinkpad_acpi
thinkpad_acpi: ThinkPad ACPI Extras v0.18-20071013
thinkpad_acpi: <a class="moz-txt-link-freetext" href="http://ibm-acpi.sf.net/">http://ibm-acpi.sf.net/</a>
thinkpad_acpi: ThinkPad BIOS 1RETDRWW (3.23 ), EC 1RHT71WW-3.04
thinkpad_acpi: IBM ThinkPad T41
leon@shpritsz:~$

I'll report back if it happens again.

Thanks, Leon.

  
      
          </pre>
        </blockquote>
      </blockquote>
      <pre wrap="">  
      </pre>
    </blockquote>
  </blockquote>
  <pre wrap=""><!---->
  </pre>
</blockquote>
</body>
</html>

--------------020500080801060302050801--