Method for direct memory access (DMA) to access data

Abstract

The invention provides an improved method for direct memory access (DMA) to access data, which comprises the following steps of: when a controller corresponding to a high-speed interface is initialized, allocating m direct memory access (DMA) hardware descriptors for the controller, and allocating n data buffer units for a slave data buffer part of the controller; constructing the software descriptor of the buffer part into a two-dimensional array enabling each row of the software descriptor to point to one buffer unit, wherein the value of a first subscript of each array item in the array ranges from 0 to n-1, and the value of a second subscript of each array item ranges from 0 to 1; respectively storing the virtual address and the physical address of the same buffer unit in a column of the two-dimensional array of the corresponding software descriptor; creating an index to establish a guidance relationship between the buffer unit and the software descriptor, the index including an idle index and a scheduling index; and achieving access of the CPU and the DMA to the memory in the data transmission process of the high-speed interface at least based on the index, the hardware descriptor and the software descriptor.

Description

technical field [0001] The present invention relates to direct memory access DMA technology, more specifically, relates to the technology of DMA accessing data. Background technique [0002] High-speed interfaces such as PCIe, USB, and Ethernet basically use DMA combined with interrupt mode when accessing memory during the communication sending and receiving process, and traditional streaming DMA mapping is used for data buffering. In this way, each receiving and sending will involve the application and release of the data buffer memory, which affects the speed of receiving and sending, and increases the power consumption of the hardware. At the same time, there is also the problem of cache consistency. It is necessary to call the API to refresh or invalidate the cache cache, which makes cache cache maintenance time-consuming, affects the performance of high-speed interface transmission and reception, and increases hardware power consumption. Therefore, it is necessary to p...

AI Technical Summary

Problems solved by technology

In this way, each receiving and sending will involve the application and release of the data buffer memory, affecting the speed of receiving and sending, and increasing the power consumption of the hardware

At the same time, there is also the problem of cache consistency. It is necessary to call the API to refresh or invalidate the cache cache, which makes cache cache maintenance time-consuming, affects the performance of high-speed interface transmission and reception, and increases hardware power consumption.

Images