loadunload

Loadunload refers to the process of loading a resource into memory for use by a program and unloading it when it is no longer needed. In computing, this typically applies to software modules such as dynamic libraries, plugins, and device drivers, but it can also apply to kernel modules and other loadable resources. The action is managed by a loader or runtime environment that controls when code and data are brought into or removed from memory.

During loading, the loader resolves dependencies, maps the module into the process or kernel address space, performs relocation if necessary, and invokes an initialization entry point. Unloading involves performing cleanup, calling finalization routines, and freeing resources. It often requires that no other code is currently using the module’s symbols or that references are properly tracked, otherwise unloading can be unsafe or blocked.

Advantages of loadunload include reduced memory footprint, shorter startup times, and the ability to extend functionality at runtime without restarting the application. It also enables on-demand loading of features, improving responsiveness and modularity. Risks include instability or crashes if modules are unloaded while still in use, symbol conflicts, version and ABI incompatibilities, and potential security vulnerabilities from loading untrusted code. Effective management typically uses reference counting, mutual exclusion, version checks, and careful lifecycle control to ensure safe loading and unloading.

Common examples appear in operating systems and programming languages. Linux uses kernel modules that can be inserted and removed, Windows provides dynamic libraries that can be loaded and freed, and many languages support dynamic or runtime loading of plugins and libraries. Related concepts include dynamic linking, hot-swapping, and plugin architectures.