Re: wrong ACPI estimate (fwd)

Re: wrong ACPI estimate (fwd)

[Len Brown]

Alexey, Henrique,
Is this related to the power supply sysfs wrong-units bug?

thanks,
Len Brown, Intel Open Source Technology Center

---------- Forwarded message ----------
Date: Thu, 15 Jan 2009 10:14:41 -0800
From: "Kok, Auke" <auke-jan.h.kok@intel.com>
To: Winfried Tilanus <winfried@tilanus.com>
Cc: "power@bughost.org" <power@bughost.org>
Subject: Re: wrong ACPI estimate

Winfried Tilanus wrote:
> Hi,
> 
> I am running Debian Lenny on a thinkpad x61s, with tp_smapi modules. I
> incorrectly thought the following problem was fixed:
> 
> $ sudo powertop -d | grep "Power usage"
> Power usage (ACPI estimate): 200.4W (0.3 hours)
> 
> This problem is discussed in Debian bug 497929
> (http://bugs.debian.org/497929)
> The problem is fixed by reverting the a code change mentioned there:
> 
> @@ -644,7 +651,7 @@ void print_battery_sysfs(void)
>                         continue;
>                 memset(line, 0, 1024);
>                 if (fgets(line, 1024, file) != NULL) {
> -                       watts_drawn = strtoull(line, NULL, 10) / 1000000.0;
> +                       amperes_drawn = strtoull(line, NULL, 10) /
> 1000000.0;
>                 }
>                 fclose(file);
> 
> Is there a way to fix this?

I doubt that we can fix this for everyone :/

unfortunately it seems that the numbers reported back to the kernel are of
different types on various systems.

I will revert this change in the next release, just because the initial change
that did this did not take this into account, and I'd rather be back at the
initial (broken) stage than another (broken) stage.

We'll have to figure out we somehow can distinguish between the two units when we
read the data, which will likely be more complex, or involve fixing various ACPI
drivers to unify their number reporting.

Auke

_______________________________________________
Power mailing list
Power@bughost.org
http://www.bughost.org/mailman/listinfo/power

Re: wrong ACPI estimate (fwd)

[Alexey Starikovskiy]

Len Brown wrote:
> Alexey, Henrique,
> Is this related to the power supply sysfs wrong-units bug?
Yes.

Regards,
Alex.
> 
> thanks,
> Len Brown, Intel Open Source Technology Center
> 
> ---------- Forwarded message ----------
> Date: Thu, 15 Jan 2009 10:14:41 -0800
> From: "Kok, Auke" <auke-jan.h.kok@intel.com>
> To: Winfried Tilanus <winfried@tilanus.com>
> Cc: "power@bughost.org" <power@bughost.org>
> Subject: Re: wrong ACPI estimate
> 
> Winfried Tilanus wrote:
>> Hi,
>>
>> I am running Debian Lenny on a thinkpad x61s, with tp_smapi modules. I
>> incorrectly thought the following problem was fixed:
>>
>> $ sudo powertop -d | grep "Power usage"
>> Power usage (ACPI estimate): 200.4W (0.3 hours)
>>
>> This problem is discussed in Debian bug 497929
>> (http://bugs.debian.org/497929)
>> The problem is fixed by reverting the a code change mentioned there:
>>
>> @@ -644,7 +651,7 @@ void print_battery_sysfs(void)
>>                         continue;
>>                 memset(line, 0, 1024);
>>                 if (fgets(line, 1024, file) != NULL) {
>> -                       watts_drawn = strtoull(line, NULL, 10) / 1000000.0;
>> +                       amperes_drawn = strtoull(line, NULL, 10) /
>> 1000000.0;
>>                 }
>>                 fclose(file);
>>
>> Is there a way to fix this?
> 
> I doubt that we can fix this for everyone :/
> 
> unfortunately it seems that the numbers reported back to the kernel are of
> different types on various systems.
> 
> I will revert this change in the next release, just because the initial change
> that did this did not take this into account, and I'd rather be back at the
> initial (broken) stage than another (broken) stage.
> 
> We'll have to figure out we somehow can distinguish between the two units when we
> read the data, which will likely be more complex, or involve fixing various ACPI
> drivers to unify their number reporting.
> 
> Auke
> 
> _______________________________________________
> Power mailing list
> Power@bughost.org
> http://www.bughost.org/mailman/listinfo/power