A Drive Adaptive Loading System

Abstract

The invention provides a driver-adaptive loading system. The system comprises an MCU mainboard, multiple submodules, a device arbitrator and a decoder, wherein each submodule comprises a memory and a functional device; the MCU mainboard is connected with the memory and the functional device in each submodule through a data line and is also connected with the device arbitrator and the decoder; the device arbitrator is connected with each submodule through a preset bus; the decoder is connected with the memory and the functional device in each submodule through a local bus; a driving program, a device identification code, a version code and a device type of a current device at least are stored in the memories; and a bus system capable of supporting hot plug is arranged in the MCU mainboard. The system supports hot plug and does not need to be restarted when a user adds or replaces an interface device; and besides, a device driver gets rid of binding to a bus, the user develops a standard device process, and bus information serves as parameters to dynamically control read-write of the device driver.

Description

technical field [0001] The invention relates to the field of driving technology, in particular to a driving adaptive loading system. Background technique [0002] At present, there are two types of known driver loading: 1. Static loading, that is, compiling the driver directly into the kernel. 2. Dynamic loading, that is, separate the driver from the kernel, and manually load the driver into the kernel when needed. However, these two methods currently have their own shortcomings and deficiencies. [0003] The static loading method has the following shortcomings: 1) It is inconvenient to debug, and every time the driver is modified, the kernel needs to be recompiled and downloaded, which is inefficient. 2) If there are many statically loaded drivers, the kernel capacity will be large and the storage space will be wasted. 3) Only known devices are supported, new unknown devices are not supported, and old devices cannot support newly developed devices. [0004] Although dyn...

AI Technical Summary

Problems solved by technology

However, at present, both methods have their own defects and deficiencies

[0003] The method of static loading has the following disadvantages: 1) Debugging is inconvenient, and the kernel needs to be recompiled and downloaded every time the driver is modified, which is inefficient

2) If there are many statically loaded drivers, the kernel capacity will be large and storage space will be wasted

3) Only known devices are supported, and unknown new devices are not supported, and old devices cannot support newly developed devices

[0004] Although dynamic loading overcomes the above shortcomings of static loading, it also has its own shortcomings. The specific points are as follows: 1) For mass production, the operation steps are increased, which is easy to increase the error rate, and requires on-site maintenance by engineers

2) If there are many device models, it is difficult to maintain the driver version and operating system version

3) The interface daughter card in the embedded field is tightly bound to the working bus, and the peripheral drivers are all bound to the fixed bus code, which cannot be written across manufacturers

4) Only provide support for existing equipment, and cannot form support for new devices developed in the future

It is inconvenient to obtain and upload the driver. In the embedded field, there are often no redundant communication interfaces. Even if the corresponding peripheral driver is written, it is difficult to upload to the system.

5) The current embedded boards generally do not support hot swap, that is, after changing the corresponding peripheral type, the embedded board needs to be restarted, so for some key devices that cannot be restarted, there is no way to perform single Expansion and replacement of board interface

Images