Hi All,
We have refactored/redesigned the existing measured boot driver present in the TF-A repo to support it with multiple backend driver(s) (for example, TCG Event Log, physical TPM, etc) instead of it being strongly coupled with the TCG Event Log driver. Proposed refactored patches are posted here: https://review.trustedfirmware.org/q/topic:%22refactor-mb%22+(status:open%20...) Any feedback/comments on these patches are much appreciated.
These patches mainly consist of the below changes:
1. Move image measurement in the generic layer, just after loading and authentication of the image. Previously, the platform layer was responsible for the measurement. For example, the Arm FVP platform layer was doing it as part of the post-load hook operation. 2. Measurement and recording of the images loaded by BL1. Previously, DTB config files loaded by BL1 were not part of measured at all. Also, it looks safer and cleaner approach to record the measurement taken by BL1 straightaway in TCG Event log buffer/physical TPM/any other TPM backend instead of deferring these recordings to BL2. 3. Pass Event Log buffer information from BL1 to BL2 so that the TCG Event Log buffer initialised by BL1 extended further with the measurements taken by BL2.
Note: These patches neither add any new functional backend driver for measured boot nor update any existing backend driver functionality (i.e. TCG Event Log driver). These changes only structured the measured boot code to provide a space to plug in any new backend driver(s) in future for the measured boot.
Thanks, Manish Badarkhe
Hi all,
On 9/21/21 1:46 PM, Manish Badarkhe wrote:> We have refactored/redesigned the existing measured boot driver present
in the TF-A repo to support it with multiple backend driver(s) (for example, TCG Event Log, physical TPM, etc) instead of it being strongly coupled with the TCG Event Log driver.
Proposed refactored patches are posted here: https://review.trustedfirmware.org/q/topic:%22refactor-mb%22+(status:open%20...)
Any feedback/comments on these patches are much appreciated.
Thank you for the feedback received so far.
It looks like most of the review comments have been addressed now and I think we are on the verge of completing the reviews of these patches. There are still a couple of open comments (in particular about certificates measurement) but I think they should not block these patches, as we can further discuss that topic separately on the mailing list.
I would like to target next Monday (EOD) to merge these initial patches, as they've been pending for almost 2 weeks now and the v2.6 release is approaching. If you've got any last minute comments, please make sure to post them before then.
Best regards, Sandrine
tf-a@lists.trustedfirmware.org