Module loading method, device, chip, electronic device and readable storage medium
By directly loading kernel object (ko) files in kernel space, the problem of increasing loading time in the existing technology is solved, faster module loading is achieved, and architectural constraints are maintained.
CN119739430APending Publication Date: 2025-04-01HUAWEI TECH CO LTD
0 Cites 0 Cited by
Patent Information
- Application Number
- CN202311290671.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-30
- Publication Date
- 2025-04-01
AI Technical Summary
Technical Problem
When loading kernel object (ko) files, the existing technology needs to enter from the user state to the kernel state and copy from the user state file to the kernel state, resulting in an increase in loading time and unable to meet performance requirements.
Method used
By directly loading ko files in the kernel space, the process of cutting from user mode to kernel mode and file copying is avoided, and module loading is directly completed in the kernel space.
Benefits of technology
It saves the system call overhead and file copy time required in the prior art, improves loading speed, and does not require source code or .o/.h file delivery, maintains the constraints of the existing architecture.
✦ Generated by Eureka AI based on patent content.
Abstract
The embodiment of the present application provides a module loading method, a module loading device, a chip, an electronic device and a computer-readable storage medium, wherein the module loading method is applied in the system startup process, including: after the kernel startup initialization, during the process of starting the execution of the init process or before starting the execution of the init process, the ko file is read based on the kernel space; based on the read ko file, the module loading is completed in the kernel space. The embodiment of the present application does not require source code or .o / .h delivery, does not destroy the architectural constraints of the existing ko decoupling binary delivery, and can flexibly adjust which ko to load first and which to delay loading according to the business; and can directly load the ko file in the kernel space, without reading the ko module file from the user state, resulting in an increase in the system call overhead, saving the time required in the prior art to switch from the user state to the kernel state, and then copy the user state file to the kernel state operation.
Need to check novelty before this filing date? Find Prior Art