Detlef Drewanz wrote:
> Mark Haywood schrieb, On 20.05.09 14:30:
>> Detlef Drewanz wrote:
>>> Hi,
>>> I am using b111b (Beta for OSOL2009.06) on my Toshiba Tecra M9 Laptop
>>> and facing still some issues with Power Management.
>>>
>>> BTW: I upgraded with the update-manager from osol2008.11 through
>>> dev-build 108 to b111b.
>>>
>>> 1. I had to set in /etc/power.conf
>>>     cpupm enable poll-mode
>>>    to see powertop displaying the speed-changes between 2001 Mhz and
>>> 800 Mhz.
>>>
>>> 2. Now if I suspend and resume my laptop, the system only runs with
>>> turbo mode (2001 Mhz) and will no longer change the cpu speed. So
>>> after resume I am no longer able to bring the system back into
>>> variable cpu speeds. Is this a know issue. Is there a workaround
>>> available ? Or is it just a wrong display  from powertop ? But kstat
>>> -m cpu_info also shows this speed.
>>
>> It is a bug with no workaround. See:
>>
>> http://defect.opensolaris.org/bz/show_bug.cgi?id=8636
>>
>> I'm working on a fix for it. cpupm event-mode does not have this 
>> problem.
>
> Hmm, but for me it looks that event-mode also do not power management 
> of my cpu. It runs always with trubo mode.

I thought that I recalled you having a problem with event-mode in some 
earlier email. You verified this was the case by running powertop while 
in event-mode?

>
> So if there is no workaround available, that would mean that 
> osol2009.06 would not do power management on Toshiba M9 (and possibly 
> others) after first suspend/resume. That would give a lot of upgraders 
> a bad impression against osol2008.11.

Why? Given what I have seen I don't believe that this worked correctly 
in 2008.11 either.

>
> Maybe there is some quick workaround to set with mdb some value ... ?

Not that I can think of. I believe that there might be a workaround so 
that event-mode might work on your system though. If so, then maybe 
someone can respond to this thread with that info.

Mark

>
>>
>> Mark
>>
>>>
>>> 3. After resume, I have seen that runnig prstat it shows me initially
>>> this crazy last line:
>>>
>>> Total: 78 processes, 209 lwps, load averages: 1533,45, 504,96, 182,88
>>>
>>> I have no clue why this load averages after 12 hours suspend are that
>>> high.
>>>
>>> Any comments ? To solve 2. is really important. If that will be not
>>> fixed until osol2009.06 comes out, that is for me a really big
>>> showstopper, because I do not want my laptop to run every time at full
>>> speed.
>>>
>>> BTW: I upgraded with the update-manager from osol2008.11 through
>>> dev-build 108 to b111b.
>>>
>>>
>>> Detlef
>>
>


Reply via email to