RE: regarding Intel E810

["Zhang, Qi Z" <qi.z.zhang@intel.com>]

[-- Attachment #1: Type: text/plain, Size: 1707 bytes --]

Hi Krishnakumar:

I’m not an expert on SGX, but as I know there are many limitations to running code in the SGX enclave due to security considerations. usually, the DMA should be setup outside of the SGX, so I guess you should not run rte_eal_init within SGX.

Regards
Qi

From: Sudharshan Krishnakumar <sudhar.ckrishna@gmail.com>
Sent: Saturday, March 5, 2022 5:38 AM
To: dev@dpdk.org; users@dpdk.org; Yang, Qiming <qiming.yang@intel.com>; Zhang, Qi Z <qi.z.zhang@intel.com>
Subject: regarding Intel E810

Hi All,
I am using Intel E810 with DPDK v20.11, with ethernet ports bound to DPDK(vfio-pci).
Running DPDK software libraries within a secure-enclave(Intel SGX technology) sandbox environment,
and running into an issue during rte_eal_init invocation.
During launch of DPDK application(examples/packet_ordering), when rte_eal_init gets called,
it results in a probe of ethernet devices,which invokes ICE(Intel E810 NIC) Poll-Mode-Driver(PMD) apis.

When ICE PMD driver sends command to NIC(Network Interface Card),to get firmware version,
it times out in this api-> ice_sq_send_cmd_nolock().
Question-> Can someone point me to any Intel documentation for E810,
that describes how ICE PMD(or any software driver) can communicate with E810 NIC's firmware,
like what registers to write the command, and how to check for the response from firmware.

Question-> Also, is there any other alternate way(other than ethtool), to check if firmware received the command,
and what was the status of the last command processed by firmware.
Note: Even if there is a way to check firmware command status thro ethtool,
I cant use ethtool,since ports are bound to DPDK.

Thanks,
Sudha

[-- Attachment #2: Type: text/html, Size: 4877 bytes --]