Wire-speed packet management in a multi-pipeline network processor

A method for maintaining processing order of data packets in a process flow, the method comprising:

• receiving packets in a packet input queue, each packet in said queue having a unique flow-identification and a packet order;
• providing a network processor having a plurality of pipeline units therein, each pipeline unit (PU) having an input and an output disposed within the network processor, and being capable of accepting a newly-received packet from the packet input queue and comprising a predetermined number of pipeline stages for processing the newly-received packet by the network processor, wherein the number of pipeline stages is at least two;
• associating flow-identification content addressable memories (FICAM) with the pipeline units, such that each said FICAM is associated with a respective PU and comprises a number of locations equal to the number of pipeline stages in said respective PU, for accepting respective flow-identifications of packets being processed by the pipeline stages of said respective PU;
• and
• dispatching said newly-received packet for processing in the network processor by one or more of said pipeline units responsively to a comparison between the flow-identification of the received packet and contents of said FICAM,
• said dispatching said newly-received packet including:
• storing a previously-received flow-identification in one of the locations in said FICAM, the previously-received flow-identification corresponding to a previously-received packet dispatched for processing by a pipeline unit (PU);
• and
• comparing a newly-received flow-identification corresponding to said newly-received packet with said contents of said FICAM;
• determining whether said newly-received flow-identification is same as said previously-received flow-identification;
• generating a hit message upon said determining that said newly-received flow-identification is the same as said previously-received flow-identification;
• and
• generating a miss message if no comparison of any FICAM generates a hit message;
• and
• upon generating a hit message, rescheduling said newly-received packet such that said newly-received packet is not placed in a PU until said miss message is generated, to maintain processing order of said newly-received and said previously-received packets.