> Thanks for raising the issue. It is completely defeating the generic > ethdev API. We must not have different behaviours depending of the > driver. Why it cannot be fixed in the driver?
I should probably refer to the thread, where the concern was raised: http://dpdk.org/ml/archives/dev/2016-July/044555.html So, the issue is that i40evf *only support* CRC stripping for some setups (i40e Linux driver for PF, i40evf DPDK driver VF). The l3fwd application disables CRC stripping for all ports, which leads to i40evf_dev_configure() failure -- which from my POV is correct. Mis-configuring a port shouldn't be allowed. I'm open to suggestions here. What would be a better way to solve this? Maybe just adding a command-line option to the l3fwd application is a better way around? Bj?rn
