Method for network processor dynamically loading microcode

Abstract

The invention discloses a method for microcode dynamic loading of a network processing unit, wherein, a microcode file is compiled in advance by the method and the microcode file comprises more than one function microcode while each function microcode corresponds to a specific processing function; during the operation of the network processing unit, a function microcode, corresponding to a corresponding microengine, is exacted from the microcode file according to the adoption condition of the current business, and then loaded to the corresponding microengine. By utilizing the method of the invention, a limited instruction space of the network processing unit can be fully utilized. When the operational environment of equipment is changed, the microcode in the instruction space of the network processing unit can be dynamically updated in the operation process of the equipment so as to provide the corresponding function.

Description

technical field [0001] The invention relates to network communication technology, in particular to a method for dynamically loading microcodes by a network processor. Background technique [0002] In the current field of network communication, with the rapid development of mobile multimedia subsystem (IMS) and third-generation (3G) communication technology, the emergence of open next-generation network technology based on IP, various new protocols and new services emerge in an endless stream , the expansion of the network scale, and the rapid increase of users' demand for bandwidth. For equipment providers, this poses serious challenges, namely: how to quickly and flexibly launch support for new services, how to ensure that service quality is not reduced, how to improve product personalization, and so on. [0003] Although traditional network equipment based on a single CPU software architecture has advantages in quickly launching new services and supporting new protocols, ...

AI Technical Summary

Problems solved by technology

However, the disadvantages of this processing method are also obvious: developers with experience in network processor development know that the microcode operation in the network processor has a strict timing arrangement, and the biggest requirement for it is to be fast and efficient. For forwarding, table lookup, and business processing, there are very few codes in the microcode that are not sensitive to its forwarding and business processing capabilities. For the microcode of the network processor, it is impossible to tolerate a certain piece of code from the memory during operation. Loading into the reserved space for execution and then entering the normal business process will disrupt the timing of microcode operation and seriously reduce the forwarding performance of microcode. At the same time, this method cannot save space. On the contrary, in order to realize this set of loading Mechanism, but also add a lot of related mechanism implementation code, the reserved space is also a waste in disguise

Images