On Sat, Feb 20, 2016 at 05:07:08PM -0800, Stephen Hemminger wrote: > Is anyone working on making KNI conform to current kernel best practices > and attempt to get it upstream? After 2 minute code review I already > see lots and lots of things that need work.
Hi Stephen, I am working on not KNI, but on KNI enhancements: KCP and KDP. KCP: Kernel Control Path http://dpdk.org/dev/patchwork/patch/10481/ KDP: Kernel Data Path http://dpdk.org/dev/patchwork/patch/10667/ Both are based on KNI, and KNI responsibilities split into these two modules, if these cover customer requirements, they can replace current KNI. KCP: Common Linux tools support without kernel drivers inside KNI required. KDP: Slow data path communication as PMD, supports both KNI FIFO and tap device. And in mail list, problems of using out-of-tree kernel module mentioned a few times, I can take over the task to submit a generic solution to Linux upstream, as a follow up to KCP and KDP tasks, but I need some kind of experts feedback about feasibility of this work. That is why I have sent a call to mail list about this issue: http://dpdk.org/ml/archives/dev/2016-February/033279.html Thanks to Aaron, he provided his support. I would like to get your idea too. Specially I am planning to use rtnetlink and as iproute2 maintainer I believe you already have some answers I am looking for. Thanks, ferruh
