Still high on the goods of act_ctinfo being accepted into kernel net-next I thought it would be a good idea to offer the user space tc side of things to control the wee beastie.
It's a pretty shameless copy/shunt/paste/hack of m_connmark and some of the other tc action control programs and does fundamentally work (used for testing of the kernel module) There are no doubt some sharp edges that I'd like to knock the corners off before a realistic attempt at submitting, so this is a general RFC call to see what I've fundamentally done wrong and can clean up. eg. I do update uapi/linux/pkt_cls.h to match the relevant bits of what has gone upstream for act_ctinfo. Should this be a separate commit? Kevin Darbyshire-Bryant (1): tc: add support for act ctinfo include/uapi/linux/pkt_cls.h | 3 +- include/uapi/linux/tc_act/tc_ctinfo.h | 34 ++++ tc/Makefile | 1 + tc/m_ctinfo.c | 244 ++++++++++++++++++++++++++ 4 files changed, 281 insertions(+), 1 deletion(-) create mode 100644 include/uapi/linux/tc_act/tc_ctinfo.h create mode 100644 tc/m_ctinfo.c -- 2.20.1 (Apple Git-117)
