Re: [RESEND] [PATCH v2 2/2] OvmfPkg/ResetVector: Exclude SEV launch secrets page from pre-validation

["Lendacky, Thomas" <thomas.lendacky@amd.com>]

On 2/23/23 09:04, Dov Murik wrote:
> 
> 
> On 23/02/2023 16:58, Dov Murik wrote:
>>
>>
>> On 21/02/2023 11:38, Gerd Hoffmann wrote:
>>> On Mon, Feb 20, 2023 at 08:44:23AM -0600, Tom Lendacky wrote:
>>>> On 2/20/23 02:49, Dov Murik wrote:
>>>>> In order to allow the VMM (such as QEMU) to add a page with hashes of
>>>>> kernel/initrd/cmdline for measured direct boot on SNP, this page must
>>>>> not be part of the SNP metadata list reported to the VMM.
>>>>>
>>>>> Check if that page is defined; if it is, skip it in the metadata list.
>>>>> In such case, VMM should fill the page with the hashes content, or
>>>>> explicitly update it as a zero page (if kernel hashes are not used).
>>>>
>>>> Would it be better to define a new section type (similar to what I did in
>>>> the SVSM PoC)? This way, it remains listed in the metadata and allows the
>>>> VMM to detect it and decide how to handle it.
>>>
>>> Explicitly describing things sounds better to me too.
>>>
>>
>> Thanks for the feedback Tom and Gerd.
>>
>>
>> I can define a new section type OVMF_SECTION_TYPE_KERNEL_HASHES. In the AmdSev
>> target it'll cover the single MEMFD page at 00F000 (after the CPUID page).
>> Now there's a question for the QEMU side -- should QEMU then fill the page
>> and encrypt it (launch_update type=NORMAL)? (currently the whole hashes table
>> creation and encryption is done elsewhere there)

Yes, I think that is the way to go. Allocate a page in Qemu, zero it out, 
fill in the hash values at the proper location and then do a launch update 
for type NORMAL page. You can use the section type to identify the data 
you need to retrieve and encrypt.

>>
>> And on regular OvmfX64 builds - should that area should be with type
>> OVMF_SECTION_TYPE_SNP_SEC_MEM which is accepted as a type=ZERO page ?
>>
>>
>> Playing with this idea, the metadata list will add:
>>
>>
>> ; Kernel hashes section for measured direct boot
>> %define OVMF_SECTION_TYPE_KERNEL_HASHES   0x5
>>
>> ...
>>
>> ; Kernel hashes for measured direct boot, or zero page if
>> ; there are no kernel hashes / SEV secrets
>> SevSnpKernelHashes:
>>    DD  SEV_SNP_KERNEL_HASHES_BASE
>>    DD  SEV_SNP_KERNEL_HASHES_SIZE
>>    DD  SEV_SNP_KERNEL_HASHES_TYPE
>>
> 
> Or maybe this metadata section ^^^^^ should be added only if the Pcd for
> secrets+hashes page is defined?

That would be optimal if you could do that.

Thanks,
Tom

> 
> -Dov
> 
>>
>>
>> and the base/size/type of that region are defined in an
>> %if statement in ResetVector.nasmb:
>>
>>
>> %if (FixedPcdGet32 (PcdSevLaunchSecretBase) > 0)
>>    ; There's a reserved page for SEV secrets and hashes; the VMM will fill and
>>    ; validate the page
>>    %define SEV_SNP_KERNEL_HASHES_TYPE    OVMF_SECTION_TYPE_KERNEL_HASHES
>>    %define SEV_SNP_KERNEL_HASHES_BASE    (FixedPcdGet32 (PcdSevLaunchSecretBase))
>> %else
>>    ; No SEV secrets and hashes page; the VMM will validate it as another zero page
>>    %define SEV_SNP_KERNEL_HASHES_TYPE    OVMF_SECTION_TYPE_SNP_SEC_MEM
>>    %define SEV_SNP_KERNEL_HASHES_BASE    (CPUID_BASE + CPUID_SIZE)
>> %endif
>> %define SEV_SNP_KERNEL_HASHES_SIZE    (FixedPcdGet32 (PcdOvmfSecPeiTempRamBase) - SEV_SNP_KERNEL_HASHES_BASE)
>>
>>
>> (I still need to figure out the point about QEMU above.)
>>
>>
>> Is that what you had in mind?
>>
>> -Dov