[Skiboot] [PATCH] platform/witherspoon: Enable eSEL logging

[Stewart Smith]

Patrick Williams <patrick at stwcx.xyz> writes:
> On Mon, Jul 17, 2017 at 06:35:46PM +1000, Stewart Smith wrote:
>> Vasant Hegde <hegdevasant at linux.vnet.ibm.com> writes:
>
>> >> I'd prefer to have an astbmc_openbmc type that would populate data
>> >> accordingly.
>> >>
>> >> Otherwise, with this patch, on boot, we'd get a notification for
>> >> witherspoon that the BMC type is "AMI", which is (of course) incorrect.
>
> What are you going to do for machines that can optionally be one of two
> different BMC stacks?
>
>> >>
>> >
>> > Stewart,
>> >
>> > Agree. Something like below looks good?
>> >
>> > +++ b/platforms/astbmc/common.c
>> > @@ -414,3 +414,9 @@ const struct bmc_platform astbmc_ami = {
>> >          .ipmi_oem_partial_add_esel   = IPMI_CODE(0x32, 0xf0),
>> >          .ipmi_oem_pnor_access_status = IPMI_CODE(0x3a, 0x07),
>> >   };
>> > +
>> > +const struct bmc_platform astbmc_openbmc = {
>> > +       .name = "OpenBMC",
>> > +       .ipmi_oem_partial_add_esel   = IPMI_CODE(0x32, 0xf0),
>> > +       .ipmi_oem_pnor_access_status = IPMI_CODE(0x3a, 0x07),
>> > +};
>> 
>> Yeah, looks good - but are those values documented anywhere for OpenBMC?
>
> We currently just copied the expected commands from Hostboot, so they
> are not documented explicitly other than in code:
>
> https://github.com/openbmc/openpower-host-ipmi-oem/blob/master/oemhandler.hpp#L12
> https://github.com/openbmc/openpower-host-ipmi-oem/blob/master/oemhandler.cpp#L142

Okay. Looks like only the partial add esel support is implemented. I'll
grab the v2 of the patch and take out the pnor_access bit and then it
should be fine.

> Having said that, I would like us to move OpenBMC OEM commands to be
> the IANA-style OEM numbering scheme.  Does skiboot support those kind of
> commands?  I suspect by the IPMI_CODE macro only having two parameters
> the answer is no.
>
> (See the IPMI spec for netfn 2E/2F if you don't know what I'm referring
> to here.)

(after a 2 second reading of that bit of the spec), I wouldn't have a
problem with transitioning to that skiboot side. The BMC abstraction in
skiboot was put together fairly hastily in order to support p8dsu SMC
BMC stack over just having supprot for the AMI BMC stack.

-- 
Stewart Smith
OPAL Architect, IBM.