Skip to contentSkip to conversation

SVSM Development Call April 3rd, 2024

DISCUSSION Jörg Rödel 4 participants 6 messages Apr 3, 2024 – Apr 4, 2024

lore
Share & tools
lore
#0 lore raw

Hi,

Here is the usual call for agenda items for this weeks SVSM development call. My agenda so far is:

* vTPM merge fallout
* Security issue mitigations

The items we can not cover during the meeting will be deferred to the next meeting.

Please note that Europe is now also on summer time, so the meeting will be at its usual time spot again. The definitive time is always in the Google calendar I shared below.

Details of the meeting (GMeet and Calendar links, meeting time) can be found in our governance repository at:

https://github.com/coconut-svsm/governance

The meeting will be recorded and the recording eventually published.

See you all there.

Regards,

Jörg

#1 lore raw

Hi Jorg
Probably you already notice that I have raised the question on vTPM interface in coconut-svsm.

I have explained how we did the POC in a secure way for TD-partitioning with TPM CRB buffer.
Jeremi Piotrowski also explained the mechanism to expose vTPM through TPM CRB MMIO.
In Intel, we still don't understand why SVSM must use a specific vTPM protocol, but not use TCG defined TPM2 CRB buffer directly.

Unfortunately, it is hard for me to use "Google Meet", and probably I am not able to join. If possible, I hope we can continue the vTPM interface discussion in email before we make decision.

Thank you
Yao, Jiewen

Quoted 33 lines

-----Original Message-----
From: Jörg Rödel <[email protected]>
Sent: Wednesday, April 3, 2024 7:20 PM
To: [email protected]; [email protected]
Subject: SVSM Development Call April 3rd, 2024

Hi,

Here is the usual call for agenda items for this weeks SVSM development call. My agenda so far is:

* vTPM merge fallout
* Security issue mitigations

The items we can not cover during the meeting will be deferred to the next meeting.

Please note that Europe is now also on summer time, so the meeting will be at its usual time spot again. The definitive time is always in the Google calendar I shared below.

Details of the meeting (GMeet and Calendar links, meeting time) can be found in our governance repository at:

https://github.com/coconut-svsm/governance

The meeting will be recorded and the recording eventually published.

See you all there.

Regards,

Jörg

#2 lore raw
Quoted 10 lines · Jiewen

Hi Jorg
Probably you already notice that I have raised the question on vTPM interface in coconut-svsm.

I have explained how we did the POC in a secure way for TD- partitioning with TPM CRB buffer.
Jeremi Piotrowski also explained the mechanism to expose vTPM through TPM CRB MMIO. In Intel, we still don't understand why SVSM must use a specific vTPM protocol, but not use TCG defined TPM2 CRB buffer directly.

Nobody has said *must*. The enlightened interface was designed after we had all the trouble with the CRB one outlined in the previous email thread, so that's what this pull request has. However, this being open source nothing prevents you submitting a pull request for a functioning CRB interface if that's what you want.

James

#3 lore raw

Thank you to confirm that James.

Yes, I will sync internally to understand the process for submitting patch to coconut-svsm.

Thank you
Yao, Jiewen

Quoted 26 lines

-----Original Message-----
From: James Bottomley <[email protected]>
Sent: Wednesday, April 3, 2024 8:19 PM
To: Yao, Jiewen <[email protected]>; Jörg Rödel <[email protected]>; svsm- [email protected]; [email protected]
Subject: Re: SVSM Development Call April 3rd, 2024

Quoted 10 lines · Jiewen

Hi Jorg
Probably you already notice that I have raised the question on vTPM interface in coconut-svsm.

I have explained how we did the POC in a secure way for TD- partitioning with TPM CRB buffer.
Jeremi Piotrowski also explained the mechanism to expose vTPM through TPM CRB MMIO. In Intel, we still don't understand why SVSM must use a specific vTPM protocol, but not use TCG defined TPM2 CRB buffer directly.

Nobody has said *must*. The enlightened interface was designed after we had all the trouble with the CRB one outlined in the previous email thread, so that's what this pull request has. However, this being open source nothing prevents you submitting a pull request for a functioning CRB interface if that's what you want.

James

#4 lore raw
In reply to Jörg Rödel
Quoted 8 lines · Jörg Rödel

Hi,

Here is the usual call for agenda items for this weeks SVSM development call. My agenda so far is:

* vTPM merge fallout
* Security issue mitigations

As I mentioned during the call, I started a shared document to discuss early attestation architecture in SVSM:
https://docs.google.com/document/d/11ZsxP8jsviP3ddp9Hrn0rf6inttNw_Pbnz0psXlxlPs

To get write access, please send me a request.
The document is really at an early stage. As Jörg suggested, we're keeping it on gdoc for now to collaborate, but as soon as we have an agreement, we'll move to github in some markdown files.

As I mentioned, from next week (April 9th to be precise) I will be away until the end of the month. I will be back online on May 3rd. Sorry in advance for the delay in my responses.

Thanks,
Stefano

Quoted 25 lines

The items we can not cover during the meeting will be deferred to the next meeting.

Please note that Europe is now also on summer time, so the meeting will be at its usual time spot again. The definitive time is always in the Google calendar I shared below.

Details of the meeting (GMeet and Calendar links, meeting time) can be found in our governance repository at:

https://github.com/coconut-svsm/governance

The meeting will be recorded and the recording eventually published.

See you all there.

Regards,

Jörg
--
Svsm-devel mailing list
[email protected]
https://mail.8bytes.org/cgi-bin/mailman/listinfo/svsm-devel