[External] Power-on monitor for phosphor-hwmon question

James Feist james.feist at linux.intel.com
Wed Oct 23 02:51:07 AEDT 2019


On 10/22/19 6:23 AM, Matt Spinler wrote:
> 
> 
> On 10/21/2019 6:02 AM, Ivan Li11 wrote:
>>
>> Hi Team,
>>
>> Does anyone have suggestion for this ?
>>
> Hi, you're right, it doesn't support that yet.  We've had a few internal
> discussions about how to handle that gracefully, though our platforms 
> don't use
> sel-logger, but nothing solid came out of them and we don't have anything
> scheduled yet.
> 
> It would be easy enough to have a watch on the PGOOD property, and then 
> do you
> what you want, like maybe turning off the thresholds.
> 
> I heard that the d-bus sensors repository has some sort of support for 
> this,
> but since we don't use it I'm not sure what all that entails. I think it 
> may
> also force you to replace phosphor-inventor-manager with entity-manager.

It should force you to switch 100%, but you would at least have to run both.

Here's what dbus-sensors does:

https://github.com/openbmc/dbus-sensors/blob/432d1edf7ac86f69558273307a59e4b1cf86b8a6/src/Utils.cpp#L141

Basically just a dbus-match that sets a "powerStatusOn" bool that can be 
queried. Depending on the sensor type thresholds don't get crossed if 
that bool is set. We also have a similar bool for bios post, as some 
sensors come up later after power on.

> 
>> Thanks your great support in advance.
>>
>> *From:*openbmc <openbmc-bounces+rli11=lenovo.com at lists.ozlabs.org> *On 
>> Behalf Of *Ivan Li11
>> *Sent:* Friday, October 18, 2019 5:59 PM
>> *To:* openbmc at lists.ozlabs.org
>> *Subject:* [External] Power-on monitor for phosphor-hwmon question
>>
>> Hi Team,
>>
>> We found that there’s no power-on monitor mechanism in phosphor-hwmon.
>>
>> https://github.com/openbmc/phosphor-hwmon
>>
>> And it will cause incorrect threshold logs to be added to journal log 
>> by ‘sel-logger’ when system is in power off(S5) state.
>>
>> Is there any plan to implement it? Or other repository we can refer to ?
>>
>> Thanks,
>>
>> Ivan
>>
> 


More information about the openbmc mailing list