Queue scheduling method and device based on linked list

Abstract

The present invention discloses a method for scheduling queues based on a chained list. The method includes the following steps: setting the number of addresses in a queuing chained list not less than the number of queues, and partitioning the queuing chained list into different queuing sub-chained lists according to the priorities of the queues, wherein the number of the addresses in the queuing sub-chained list is not less than the total number of the queues whose priorities correspond to the queuing sub-chained list; setting for each queue a queuing chained list identifier identifying whether the each queue has queued in the queuing chained list; determining whether a queue satisfying queuing criteria has queued in the queuing chained list according to the queuing chained list identifier of the queue before the queue is added to the queuing chained list; if it has queued, adding is not processed, otherwise, the queue is added to the end of the queuing sub-chained list which corresponds to the priority of the queue, and the queuing chained list identifier of the queue is modified to an identifier identifying the queue has already queued in the queuing chained list. The present invention also discloses a device for scheduling queues based on a chained list. The present invention ensures impartiality when queues having the same priority are scheduled.

Description

technical field [0001] The present invention relates to queue scheduling technology, in particular to a queue scheduling method and device based on a linked list mixed with multiple scheduling technologies. Background technique [0002] Queue scheduling, that is, the queue passes certain trigger checks (such as receiving authorization triggers, etc.), arbitrates the queue exit conditions, and schedules the queue to exit the queue in an orderly manner according to certain rules. Commonly used scheduling algorithms mainly include round robin (RR, Round Robin) scheduling algorithm and strict priority (SP, Strict Priority) scheduling algorithm. [0003] The implementation principle of the round-robin scheduling algorithm is to poll the queue one by one and circularly to dequeue in a certain order. Schedule a queue to be dequeued every clock cycle, and then poll the next queue for dequeuing on the next clock cycle. The strict priority scheduling algorithm is to set different pr...

AI Technical Summary

Problems solved by technology

This does not reflect the idea that the same priority queue should be polled fairly, resulting in unfair dequeue

Another problem is that when there are many queues, there is a risk of FIFO overflowing. To avoid this phenomenon, it is necessary to configure a larger FIFO queue, which will consume a considerable amount of RAM resources.

If the number of queues and queue priorities are large, for example, the number of queues is 32K and the priority of queues is 6. In order to avoid overflow, at least 6 FIFOs with a depth of 32K are required for dequeue queuing, which will undoubtedly cost There are a lot of RAM resources, but a large part of the RAM resources configured above cannot be used, which also causes a serious waste of resources

Images