Use of EdgeLock Enclave "Commit" Command and potential (undocumented) side-effects

取消
显示结果 
显示  仅  | 搜索替代 
您的意思是: 
已解决

Use of EdgeLock Enclave "Commit" Command and potential (undocumented) side-effects

跳至解决方案
294 次查看
gpoulios
Contributor II

Hello NXP TechSupport team,

It is my understanding of the manuals that the ELE "Commit" command (IMX93ELEAPI, par. 3.15) is meant to be issued after successful authentication upon each boot, even though this is not explicitly stated.

Doing so, as one might imagine, will result in multiple "Commit" commands being issued for the same FW version and SRK revocation values (practically nothing new to commit to in the fuses) across the lifetime of the device.

Can the NXP TechSupport team please confirm that issuing multiple "Commit" commands (for both OEM FW version and SRK revocations, i.e. bitmask 0x10 | 0x20) when the FW version and SRK revocation bits are the same, does not have any unintended side-effects with the fuses? Would such practice result in any unnecessary (potentially irreversible) fuse blowing?

Again, my understanding is that for the same FW version, and SRK revocation list, the "Commit" command should be a no-op; I would just like this to be confirmed by NXP, given the irreversibility of fuse operations.

Thank you in advance.

Regards,

George

0 项奖励
回复
1 解答
238 次查看
Harvey021
NXP TechSupport
NXP TechSupport

Hi @gpoulios 

Your understanding is correct. We have some tests for both commits without issue。

 

Best regards

Harvey

在原帖中查看解决方案

0 项奖励
回复
2 回复数
223 次查看
gpoulios
Contributor II

Thank you @Harvey021.

0 项奖励
回复
239 次查看
Harvey021
NXP TechSupport
NXP TechSupport

Hi @gpoulios 

Your understanding is correct. We have some tests for both commits without issue。

 

Best regards

Harvey

0 项奖励
回复