Re: [PATCH 09/19] eeepc-laptop: support for super hybrid engine (SHE)

[Alan Jenkins <alan-jenkins@tuffmail.co.uk>]

Corentin Chary wrote:
> On Mon, May 25, 2009 at 12:07 PM, Alan Jenkins
> <alan-jenkins@tuffmail.co.uk> wrote:
>   
>> Corentin Chary wrote:
>>     
>>> On Mon, May 25, 2009 at 11:43 AM, Alan Jenkins
>>> <alan-jenkins@tuffmail.co.uk> wrote:
>>>
>>>       
>>>> Corentin Chary wrote:
>>>>
>>>>         
>>>>> On Sunday 24 May 2009 19:29:37 Alan Jenkins wrote:
>>>>>
>>>>>
>>>>>           
>>>>>> Corentin Chary wrote:
>>>>>>
>>>>>>
>>>>>>             
>>>>>>> On Sun, May 24, 2009 at 11:28 AM, Alan Jenkins
>>>>>>>
>>>>>>> <sourcejedi.lkml@googlemail.com> wrote:
>>>>>>>
>>>>>>>
>>>>>>>               
>>>>>>>> On 5/16/09, Len Brown <lenb@kernel.org> wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>>>>>> From: Grigori Goronzy <greg@chown.ath.cx>
>>>>>>>>>
>>>>>>>>> The older eeepc-acpi driver allowed to control the SHE performance
>>>>>>>>> preset through a ACPI function for just this purpose. SHE underclocks
>>>>>>>>> and undervolts the FSB and undervolts the CPU (at preset 2,
>>>>>>>>> "powersave"), or slightly overclocks the CPU (at preset 0,
>>>>>>>>> "performance"). Preset 1 is the default setting with default clocks and
>>>>>>>>> voltage.
>>>>>>>>>
>>>>>>>>> The new eeepc-laptop driver doesn't support it anymore.
>>>>>>>>> The attached patch adds support for it to eeepc-laptop. It's very
>>>>>>>>> straight-forward and almost trivial.
>>>>>>>>>
>>>>>>>>> Signed-off-by: Grigori Goronzy <greg@chown.ath.cx>
>>>>>>>>> Signed-off-by: Corentin Chary <corentincj@iksaif.net>
>>>>>>>>> Signed-off-by: Len Brown <len.brown@intel.com>
>>>>>>>>> ---
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>                   
>>>>>>>> Hi, out of curiosity I tried this on my EeePC 701.  I upgraded the
>>>>>>>> BIOS to the latest version available a few months ago.
>>>>>>>>
>>>>>>>> I find that the file is present and can successfully be read from.
>>>>>>>> The file returns the value "513".  If I write "1" to it, nothing
>>>>>>>> happens.  If I write "0" to it, the speakers start hissing and the
>>>>>>>> file then returns the value "512".  Writing "1" again gets it back to
>>>>>>>> normal.  There is no apparent effect on performance.
>>>>>>>>
>>>>>>>> This is stupid, because we _do_ appear to check the BIOS supported
>>>>>>>> features bitmask, but that's Asus firmware for you.  Can you please
>>>>>>>> add an extra test, so this file only allows  reads or writes if the
>>>>>>>> current value is 0 or 1?  If you're quick you might slip it into -rc8
>>>>>>>>
>>>>>>>>
>>>>>>>>                 
>>>>> Hi, Can you try this patch ? It seems to works for me.
>>>>>
>>>>>
>>>>>           
>>>> Thanks, it does make the interface less confusing. The behaviour (no
>>>> performance change, hissing speakers) is the same.
>>>>
>>>>         
>>> It works on mine (original bios). But I don't know how to see if there
>>> is a performance change.
>>> Is there a quick cpu bench ?
>>>
>>>       
>> I used:
>>
>> time for {1..10000}; do echo -n; done
>>
>> It's a bit bogus - I expect it would show if my 630Mhz processor jumped
>> to 900Mhz, but smaller changes might be lost in noise.
>>
>> <http://pavelmachek.livejournal.com/77425.html> suggests "time factor
>> $[65863223*65863159]", which should be better.
>>
>> I think it's also significant that the current (630Mhz) setting is "1".
>> I would expect "0" to be slower - but in the original 701 BIOS, 630Mhz
>> is the slower of the two speeds, right?
>>     
>
> 1 - time factor: ~ 1.574s  - default, seems to be 630Mhz
> 0 - time factor: ~ 1.01s    - seems to be 900
>
>   

How illogical :-).  Oh - I should have read the commit message, this is 
the expected order (and proper SHE just has the extra state: 2 / 
"performance").

Perhaps we should DMI-blacklist 701s with newer BIOS versions, so we 
only provide the performance control when it is available from the BIOS 
setup screen.  The specific version is well-documented e.g. on 
forum.eeeuser.com.

Thanks for your time
Alan