[etherlab-dev] Userspace fork of Etherlab

[Armin Steinhoff]

Florian Pose schrieb:
> Hallo Frank,
>
> [ clip]
>
> As long as the generic driver is not completely realtime capable and the
> kernel interfaces shall be maintained, there are strong arguments to
> keep a significant part of the master code in kernel-space.

I'm not conviienced that a significant part of the master code must be 
in the kernel space!
I did recently an integration of the openPowerlink stack and the 
userspace driver of Intel's i210 board as used within the DPDK project.
It is working in real-time under PREEMPT_RT LINUX after solving problems 
with the different buffer systems.

A similar solution should also be possible with EtherCAT ..

Cheers

Armin Steinhoff