Re: [dmidecode] [PATCH v2 1/1] dmioem: HPE OEM Record 199

[Jean Delvare]

Hi Jerry,

I'll send you all the relevant binary dump files. As a heads up though,
this is not an isolated case. In fact I do not have any example where
the CPU IDs of record type 199 make sense for AMD CPUs.

Jean

On Tue, 9 Mar 2021 09:32:10 -0700, Jerry Hoemann wrote:
> I'm not seeing this particular decode on the samples that I have
> tested with.  Looks like you may have picked up some samples that
> I don't have.  (note, I am not seeing an AMD EPYC 7451 on my tests..)
> 
> Could you email me the binary dumps whose decode you question?
> 
> thanks
> 
> Jerry
> 
> > 
> > I like it very much, thanks.
> > 
> > I tested it on my samples. The output makes sense for all Intel-based
> > systems. However for AMD-based systems, the CPU ID signature makes
> > little sense and doesn't match the installed CPU model. For example:
> > 
> > Handle 0x0003, DMI type 199, 28 bytes
> > HPE ProLiant CPU Microcode Patch Support Info
> >         CPU ID: Family 0, Model 1, Stepping 2
> >                 Date: 2018-02-09
> >                 Patch: 0x8001227
> >         CPU ID: Family 0, Model 1, Stepping 1
> >                 Date: 2017-07-14
> >                 Patch: 0x8001129
> > 
> > Handle 0x0084, DMI type 4, 48 bytes
> > Processor Information
> >         Socket Designation: Proc 1
> >         Type: Central Processor
> >         Family: Zen
> >         Manufacturer: Advanced Micro Devices, Inc.
> >         ID: 12 0F 80 00 FF FB 8B 17
> >         Signature: Family 23, Model 1, Stepping 2
> >     (...)
> >         Version: AMD EPYC 7451 24-Core Processor
> >     (...)
> > 
> > That's kind of confusing. Whatever is encoded in the CPU ID field, it
> > doesn't look like an actual CPUID number. Any idea?
> > 
> > -- 
> > Jean Delvare
> > SUSE L3 Support