cvs commit: src/usr.sbin/powerd powerd.c

[Nate Lawson]

Marcel Moolenaar wrote:
> 
> On Feb 26, 2005, at 5:15 PM, Nate Lawson wrote:
> 
>> Marcel Moolenaar wrote:
>>
>>> marcel      2005-02-27 00:23:19 UTC
>>>   FreeBSD src repository
>>>   Modified files:
>>>     usr.sbin/powerd      powerd.c   Log:
>>>   Unbreak !i386 platforms. Only i386 has APM, so don't try to fallback
>>>   on APM on other platforms.
>>>     Revision  Changes    Path
>>>   1.3       +11 -1     src/usr.sbin/powerd/powerd.c
>>
>>
>> Thanks.  It would be cleaner to move this code into apm_init_ac() and
>> acpm_get_ac() functions so the ifdef could be in just one place.
> 
> 
> Agreed. That would also make it easier to have powerd(8) on platforms
> that have neither APM nor ACPI, but have something else (possibly).

Done.  Shouldn't have drunk all that tea (yunnan gold, quite good.)

>> I think there is also one other issue.  It would be cleaner to check
>> for the acpi sysctl in the rc.d script and return immediately rather
>> than starting powerd and it bombing out once it discovers this.
> 
> 
> I'm not sure. I think PowerPC will benefit if powerd(8) grows support
> for wharever is present there. In that light, it makes sense to treat
> powerd(8) as a generic service and thus have it started unconditionally.
> 
> Just a thought,

Yep.  At least ARM has several cpufreq drivers to learn from in Linux. 
I also reread the code and saw that in the case of all AC stuff failing, 
we just use the normal policy (-n).  So this should work as-is on other 
platforms.

-- 
Nate