Re: [edk2-devel] [PATCH v2 4/4] OvmfPkg/Tcg2ConfigPei: Mark TPM MMIO range as unencrypted for SEV-ES

["Laszlo Ersek" <lersek@redhat.com>]

I need to excuse myself for two items here, where your expectation was
justified:

On 04/28/21 21:43, Tom Lendacky wrote:
> On 4/28/21 12:51 PM, Laszlo Ersek via groups.io wrote:
>> I'm going to ask for v3 after all:
>>
>> On 04/27/21 18:21, Lendacky, Thomas wrote:

>>> @@ -627,6 +627,7 @@ [Components]
>>>  
>>>  !if $(TPM_ENABLE) == TRUE
>>>    OvmfPkg/Tcg/Tcg2Config/Tcg2ConfigPei.inf
>>> +  OvmfPkg/Tcg/TpmMmioSevDecryptPei/TpmMmioSevDecryptPei.inf
>>>    SecurityPkg/Tcg/TcgPei/TcgPei.inf
>>>    SecurityPkg/Tcg/Tcg2Pei/Tcg2Pei.inf {
>>>      <LibraryClasses>
>>
>> (5) Functionally correct, but it reads more nicely (from a logical
>> dependency POV) if we place the new PEIM first.
>>
>> (Please apply to the rest of the DSC files, and the FDF files too.)
> 
> Ok, I was going with the alphabetical placement. I'll switch it up.

Well, you are not wrong; what I request for ordering between lib classes
and between FDF/DSC entries is indeed inconsistent. I guess for lib
classes, showing the construction order makes little, as that is
determined with respect to particular library instances, plus we wrangle
lib classes all the time, and keeping consistency is simplest with
alphabetical ordering. Dispatch order is *somewhat* more directly
visible in a FDF file... but yes, keeping INF references in alphabetical
order there too would certainly plausible

>>> +      DEBUG ((DEBUG_INFO, "%a: failed to map TPM MMIO address range unencrypted\n", __FUNCTION__));
>>
>> (13) Overlong line.
> 
> Ok, I'll change that. I though that was ok now since PatchCheck.py didn't
> complain.

Sorry about my pickiness; this is a point where I strongly disagree with
the rest of the edk2 maintainers -- I really do insist on 80 chars per
line, as my eyesight isn't the greatest, I totally depend on two code
windows being shown side by side, and I *also* can't work with multiple
monitors (I've tried it, I just can't). So... one monitor, mid-size
fonts, two columns of text --> 80 chars per column.

Thanks & sorry about the trouble,
Laszlo