Re: [Qemu-devel] SMBIOS (Set of 10 patches)

[Kevin O'Connor]

On Tue, Mar 11, 2014 at 11:03:06AM +0100, Gerd Hoffmann wrote:
> Issue #3: Running a diff on the dmidecode output with and without the
> patches yields this:

I think it would be best to get the patch series to the point that
there is no diff between old and new.  That will make review easier,
and subsequent patches can then add new features.

> --- dmidecode.master  2014-03-11 10:38:06.799233009 +0100
> +++ dmidecode.smbios  2014-03-11 10:39:36.664377785 +0100
> @@ -1,20 +1,20 @@
>  # dmidecode 2.12
>  SMBIOS 2.4 present.
> -10 structures occupying 304 bytes.
> -Table at 0x000F09D0.
> +10 structures occupying 351 bytes.
> +Table at 0x000F09A0.
> 
> That comes from upgrading some of the tables to newer versions, ok.
>  
>  Handle 0x0000, DMI type 0, 24 bytes
>  BIOS Information
> -     Vendor: Bochs
> -     Version: Bochs
> -     Release Date: 01/01/2011
> +     Vendor: QEMU
> +     Version: pc-i440fx-2.0
> +     Release Date: 01/01/2014
>       Address: 0xE8000
>       Runtime Size: 96 kB
>       ROM Size: 64 kB
>       Characteristics:
>               BIOS characteristics not supported
>               Targeted content distribution is supported
> -     BIOS Revision: 1.0
> +     BIOS Revision: 0.0
> 
> I think we should not generate a type0 table unless -smbios type0=... is
> explicitly specified on the qemu command line.  It is about the
> firmware, and we should leave it to the firmware to fill it by default.
> If you are running OVMF (EFI) instead of SeaBIOS you should see it in
> the dmidecode output.

Everything that SeaBIOS puts into table 0 is hard coded.  I'd prefer
it if QEMU created the table (with the same hardcoded fields) because
having split ownership of the smbios is painful.

[...]
> -Handle 0x0401, DMI type 4, 32 bytes
> +Handle 0x0400, DMI type 4, 35 bytes
>  Processor Information
> -     Socket Designation: CPU 1
> +     Socket Designation: CPU 0
> 
> Hmm?
> 
>       Type: Central Processor
>       Family: Other
> -     Manufacturer: Bochs
> +     Manufacturer: QEMU
>       ID: 63 06 00 00 FD FB 8B 07
> -     Version: Not Specified
> +     Version: pc-i440fx-2.0
>       Voltage: Unknown
>       External Clock: Unknown
> 
> Ok.
> 
> -     Max Speed: 2000 MHz
> -     Current Speed: 2000 MHz
> +     Max Speed: Unknown
> +     Current Speed: Unknown
> 
> Where does 2000 MHz come from?  Does SeaBIOS pull something out of thin
> air or does it try to measure the speed?

The 2000 is hardcoded in SeaBIOS (see smbios_init_type_4() ).

> -Handle 0x1100, DMI type 17, 21 bytes
> +Handle 0x1100, DMI type 17, 27 bytes
>  Memory Device
>       Array Handle: 0x1000
> -     Error Information Handle: 0x0003
> +     Error Information Handle: Not Provided
> 
> Same question.

I don't see the 0x0003 in smbios_init_type_17() but I think it must
come from some hardcode in seabios.

-Kevin