Re: [PATCH v2 2/2] efi/libstub: Add get_event_log() support for CC platforms

[Kuppuswamy Sathyanarayanan]

On 2/27/24 5:19 AM, Ilias Apalodimas wrote:
> On Sat, 24 Feb 2024 at 09:31, Kuppuswamy Sathyanarayanan
> <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx> wrote:
>>
>> On 2/23/24 5:24 AM, Ilias Apalodimas wrote:
>>> Apologies for the late reply,
>>>
>>>
>>> On Mon, 19 Feb 2024 at 09:34, Kuppuswamy Sathyanarayanan
>>> <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx> wrote:
>>>> Hi Ilias,
>>>>
>>>> On 2/18/24 11:03 PM, Ilias Apalodimas wrote:
>>>>> On Thu, 15 Feb 2024 at 05:02, Kuppuswamy Sathyanarayanan
>>>>> <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx> wrote:
>>>>>> To allow event log info access after boot, EFI boot stub extracts
>>>>>> the event log information and installs it in an EFI configuration
>>>>>> table. Currently, EFI boot stub only supports installation of event
>>>>>> log only for TPM 1.2 and TPM 2.0 protocols. Extend the same support
>>>>>> for CC protocol. Since CC platform also uses TCG2 format, reuse TPM2
>>>>>> support code as much as possible.
>>>>>>
>>>>>> Link: https://uefi.org/specs/UEFI/2.10/38_Confidential_Computing.html#efi-cc-measurement-protocol [1]
>>>>>> Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx>
>>>>> [...]
>>>>>
>>>>>> +void efi_retrieve_eventlog(void)
>>>>>> +{
>>>>>> +       efi_physical_addr_t log_location = 0, log_last_entry = 0;
>>>>>> +       efi_guid_t cc_guid = EFI_CC_MEASUREMENT_PROTOCOL_GUID;
>>>>>> +       efi_guid_t tpm2_guid = EFI_TCG2_PROTOCOL_GUID;
>>>>>> +       int version = EFI_TCG2_EVENT_LOG_FORMAT_TCG_2;
>>>>>> +       efi_tcg2_protocol_t *tpm2 = NULL;
>>>>>> +       efi_cc_protocol_t *cc = NULL;
>>>>>> +       efi_bool_t truncated;
>>>>>> +       efi_status_t status;
>>>>>> +
>>>>>> +       status = efi_bs_call(locate_protocol, &tpm2_guid, NULL, (void **)&tpm2);
>>>>>> +       if (status == EFI_SUCCESS) {
>>>>>> +               status = efi_call_proto(tpm2, get_event_log, version, &log_location,
>>>>>> +                                       &log_last_entry, &truncated);
>>>>>> +
>>>>>> +               if (status != EFI_SUCCESS || !log_location) {
>>>>>> +                       version = EFI_TCG2_EVENT_LOG_FORMAT_TCG_1_2;
>>>>>> +                       status = efi_call_proto(tpm2, get_event_log, version,
>>>>>> +                                               &log_location, &log_last_entry,
>>>>>> +                                               &truncated);
>>>>>> +                       if (status != EFI_SUCCESS || !log_location)
>>>>>> +                               return;
>>>>>> +               }
>>>>>> +
>>>>>> +               efi_retrieve_tcg2_eventlog(version, log_location, log_last_entry,
>>>>>> +                                          truncated);
>>>>>> +               return;
>>>>>> +       }
>>>>>> +
>>>>>> +       status = efi_bs_call(locate_protocol, &cc_guid, NULL, (void **)&cc);
>>>>>> +       if (status == EFI_SUCCESS) {
>>>>>> +               version = EFI_CC_EVENT_LOG_FORMAT_TCG_2;
>>>>>> +               status = efi_call_proto(cc, get_event_log, version, &log_location,
>>>>>> +                                       &log_last_entry, &truncated);
>>>>>> +               if (status != EFI_SUCCESS || !log_location)
>>>>>> +                       return;
>>>>>> +
>>>>>> +               efi_retrieve_tcg2_eventlog(version, log_location, log_last_entry,
>>>>>> +                                          truncated);
>>>>>> +               return;
>>>>>> +       }
>>>>>> +}
>>>>> [...]
>>>>>
>>>>> I haven't looked into CC measurements much, but do we always want to
>>>>> prioritize the tcg2 protocol? IOW if you have firmware that implements
>>>>> both, shouldn't we prefer the CC protocol for VMs?
>>>> According the UEFI specification, sec "Conidential computing", if a firmware implements
>>>> the TPM, then it should be used and CC interfaces should not be published. So I think
>>>> we should check for TPM first, if it does not exist then try for CC.
>>> Ok thanks, that makes sense. That document also says the services
>>> should be implemented on a virtual firmware.
>>> I am unsure at the moment though if it's worth checking that and
>>> reporting an error otherwise. Thoughts?
>> IMO, it is not fatal for the firmware to implement both protocols. Although, it
>> violates the specification, does it makes sense to return error and skip
>> measurements? I think for such case, we can add a warning and proceed
>> with TPM if it exists.
> If you have a TPM, the current code wouldn't even look for CC (which
> we agreed is correct).
> The question is, should we care if a firmware exposes the CC protocol,
> but isn't virtualized

AFAIK, even if a firmware improperly uses this protocol (in a non-virtual
environment), it should not be a fatal issue. So, if we add such a check,
it will be just a spec compliance check. Also, a firmware can improperly
use any existing EFI interfaces in n other ways. But, we cannot check for
all such cases, right? So personally I think it is not needed. But I am fine
either way.

If we want to add such check, I think we should either cc_platform_has()
or CPU feature flag check for it.

>
> Thanks
> /Ilias
>>> Thanks
>>> /Ilias
>>>> https://uefi.org/specs/UEFI/2.10/38_Confidential_Computing.html#confidential-computing
>>>>
>>>>> Thanks
>>>>> /Ilias
>>>> --
>>>> Sathyanarayanan Kuppuswamy
>>>> Linux Kernel Developer
>>>>
>> --
>> Sathyanarayanan Kuppuswamy
>> Linux Kernel Developer
>>
-- 
Sathyanarayanan Kuppuswamy
Linux Kernel Developer