timeout on sgx_ra_get_msg1

[Elephant]

Hi,

I am testing Remote Attestation and I am having some weird results.  My setup is an Intel Server block and a laptop running Ubuntu (both have Intel SGX CPUs) doing the remote attestation.  When I use the Intel Server block as the "Service Provider", i.e. the one who provisions the secret and the one talking to the Intel Attestation Server, my test works.  But when I make that Intel Server block as a slave for my remote attestation, whenever it calls "sgx_ra_get_msg1", I am getting timeouts with that call and when I checked the status of the "aesmd" service I get the following message "[ADMIN]EPID Provisioning failed due to network error".  I am not sure what is the issue here since that server block was able to talk to IAS when it was the Service Provider (so I am assuming there's really no network issue).  Any ideas?  Thanks!

Kind Regards,
Elephant

[Elephant]

Some additional information for the query:

1. I saw that when starting the aesmd service, I get this error: "[ADMIN]Platform Services initialization failed due to DAL"

2. I am using the following versions: linux-sgx (3699ffd Merge pull request #188 from yuyuany/heap_init) and linux-sgx-driver (03435d3 Merge pull request #41 from avchinch/master)

3. Intel Server Block (S1200SP.86B.03.01.0020.052620170737)

Thanks!

Kind Regards,
Elephant

[o_f_]

Hi,

Same error message here using sgx1.9 pre-built libraries (release).

Any recommended resolution?

[Elephant]

Hi,

I have sent a query about this on the github repository: https://github.com/01org/linux-sgx/issues/190

The solution for now is to checkout the exact commit tagged: sgx_v1.9

This is just a temporary solution though.  Still waiting for the actual fix.

Kind Regards,
Elephant