Re: Some questions about "Add option for SEL commands for Journal-based SEL entries"

["Bills, Jason M" <jason.m.bills@linux.intel.com>]

On 3/29/2019 4:00 AM, Will Liang (梁永鉉) wrote:
> Hi Jason,
> 
> 
> Sorry to bother you again.
It's not a bother. :)
> 
> After solving sensor type and event type problem for discrete 
> sensor(memory) , I encountered another issue.
>  From the code, it seems only support getting the sensor number of 
> threshold-based sensors.
> How do I get the sensor  number of discrete sensors?
> Would you have any suggestion or plan to get sensor number of discrete 
> sensors?

Unfortunately, you have hit the crux of the issue that is preventing the 
journal-based SEL from being fully adopted, which is that we don't have 
a flexible, generic way to assign sensor numbers.

If you only need discrete sensors to be assigned numbers dynamically, 
you might be able to extend the sensor mapping that is used to get the 
number here: 
https://gerrit.openbmc-project.xyz/c/openbmc/phosphor-host-ipmid/+/12951/24/sdrutils.hpp#117. 
  If you can set a path for your discrete sensors so they are included 
in the list, it may solve your issue.

We've discussed some options for a flexible, generic solution that would 
work for everyone, but I don't believe any work has been done, yet.  You 
can view the discussion here: 
https://lists.ozlabs.org/pipermail/openbmc/2019-January/014873.html and 
in the comments on the review here: 
https://gerrit.openbmc-project.xyz/c/openbmc/phosphor-host-ipmid/+/12951.

Thanks,
-Jason