[PATCH skeleton] skeleton: Add new API SystemManager::getFRUObject()

[Yi TZ Li]

Hi Vishwa,

Thanks for the comments. I've create a new dbus API based on your
suggestions.
Please have a look:
https://github.com/openbmc/skeleton/pull/62
https://github.com/openbmc/ipmi-fru-parser/pull/14/files

Thanks,
-Yi

Vishwanatha Subbanna/India/IBM wrote on 04/11/2016 05:49:54 PM:

> From: Vishwanatha Subbanna/India/IBM
> To: OpenBMC Patches <openbmc-patches at stwcx.xyz>, OpenBMC Members, Yi
> TZ Li/China/IBM at IBMCN
> Cc: openbmc at lists.ozlabs.org
> Date: 04/11/2016 05:49 PM
> Subject: Re: [PATCH skeleton] skeleton: Add new API
> SystemManager::getFRUObject()
>
> Hello Yi,
>
> Thanks for the patch. Here are my comments.
>
> "openbmc" <openbmc-bounces+vishwanath=in.ibm.com at lists.ozlabs.org>
> wrote on 11/04/2016 02:40:52 pm:
>
> > From: OpenBMC Patches <openbmc-patches at stwcx.xyz>
> > To: openbmc at lists.ozlabs.org
> > Date: 11/04/2016 02:41 pm
> > Subject: [PATCH skeleton] skeleton: Add new API
> SystemManager::getFRUObject()
> > Sent by: "openbmc" <openbmc-bounces
+vishwanath=in.ibm.com at lists.ozlabs.org>
> >
> > ipmi-fru-parser uses getObjectFromId('FRU_STR', fru_area_name) to
> > ask for the dbus object path of the 'fru_area_name'. The returned
> > value is also used to check the 'present' status of FRU. Sometimes
> > the 'fru_area_name' does not defined in Skeleton's ID_LOOKUP table.
> > system_manager will report a "ERROR: lookup not found" error message.
> >
>
>
> > This patch added an API similiar with getObjectFromId().
> > It does not report error when the 'fru_area_name' is not defined.
> > It can be used by ipmi-fru-parser to replace getObjectFromId().
> >
>
> If something is not there, it is still an error. So masking the
> error may not be a right approach.
>
>
> > Also the orignal "ERROR" message is changed to "WARNING".
> >
>
> IMO, we may not be solving the root cause here.. "WARNING", -or-
> "ERROR" would still bother whoever looking into it.
> So, I still feel that we need to make the actual change.
>
> When I initially proposed a solution to Norm to remove those ERRORs,
> he had an opinion that "something not there is not there, so its an
error".
>
> Here is how I look into it ( one of the implementations .. )
>
> 1: writefru data will make a call into a new API passing the fru id.
> 2: Skeleton code will return the available sections for that fru id.
> ( example: chassis, board / board, product etc )
> 3: writefru will use that and then make getObjectFromByteId for a
> particular fruid_<area>
>

> > Signed-off-by: Yi Li <adamliyi at msn.com>
> >
> > <!-- Reviewable:start -->
> > ---
> > This change is [<img src="https://reviewable.io/review_button.svg"
> > height="35" align="absmiddle" alt="Reviewable"/>](https://
> > reviewable.io/reviews/openbmc/skeleton/60)
> > <!-- Reviewable:end -->
> >
> >
> > https://github.com/openbmc/skeleton/pull/60
> >
> > Yi Li (1):
> >   skeleton: Add new API SystemManager::getFRUObject()
> >
> >  bin/system_manager.py | 40 +++++++++++++++++++++++++++-------------
> >  1 file changed, 27 insertions(+), 13 deletions(-)
> >
> > --
> > 2.7.1
> >
> >
> > _______________________________________________
> > openbmc mailing list
> > openbmc at lists.ozlabs.org
> > https://lists.ozlabs.org/listinfo/openbmc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ozlabs.org/pipermail/openbmc/attachments/20160415/e9af6aef/attachment-0001.html>