US 12,231,339 B2
Extension of openvswitch megaflow offloads to hardware to address hardware pipeline limitations
Namrata Limaye, Fremont, CA (US); Venkata Suresh Kumar Paruchuri, Bangalore (IN); Sridhar Samudrala, Portland, OR (US); and Kiran Patil, Portland, OR (US)
Assigned to Intel Corporation, Santa Clara, CA (US)
Filed by Intel Corporation, Santa Clara, CA (US)
Filed on Nov. 9, 2020, as Appl. No. 17/093,394.
Prior Publication US 2021/0144094 A1, May 13, 2021
Int. Cl. H04L 47/125 (2022.01); H04L 47/11 (2022.01); H04L 47/22 (2022.01); H04L 47/2441 (2022.01)
CPC H04L 47/125 (2013.01) [H04L 47/11 (2013.01); H04L 47/22 (2013.01); H04L 47/2441 (2013.01)] 15 Claims
OG exemplary drawing
 
1. A method implemented on a compute platform including a Network Interface Controller (NIC) having one or more ports and running software including virtual switch software, a Linux operating system having a kernel including a Traffic Classification (TC)-flower module and a NIC driver, comprising:
creating a new megaflow with a mask in the virtual switch software employing a subset of microflow fields for a packet;
creating a new hardware flow on the NIC employing a packet match scheme using at least a portion of the microflow fields;
receiving a first packet at a first port on the NIC;
detecting the first packet does not belong to a current hardware flow;
forwarding the packet from the NIC to the TC-flower module; and
at the TC flower module,
matching the first packet to a megaflow and forwarding the first packet to be processed using software based on a megaflow rule associated with the megaflow; and
extending a TC datapath to add a microflow rule to the NIC via an interface on the TC flower module to the NIC driver.