登录查看更多内容

How do you choose between static and dynamic linking for a C/C++ program on Linux?

由人工智能和领英社区提供技术支持

When you write a C/C++ program on Linux, you need to link it with other libraries that provide functions, data, or resources that your program needs. Linking can be done in two ways: static or dynamic. Static linking means that the library code is copied into your executable file at compile time, while dynamic linking means that the library code is loaded into memory at run time. Both methods have advantages and disadvantages, and choosing the right one depends on several factors. In this article, we will explain the differences between static and dynamic linking, and how to decide which one to use for your C/C++ program on Linux.

此文章中的业界达人

由社区从 8 条内容中精选。了解更多

Sudeep K S

Firmware Engineer @ Intel Corporation | Vision IP Development | Camera Driver Development
Vikram vel

Firmware Engineer
Christian Huaux

Senior Developer & Technical Architect | Skilled in High-Performance Computing, Game Engine Optimization, and…

1 Benefits of static linking

Static linking has some benefits that make it attractive for certain scenarios. First, it simplifies the deployment of your program, as you don't need to worry about distributing or installing the libraries that your program depends on. Your executable file contains everything it needs to run. Second, it improves the performance of your program, as it avoids the overhead of loading and resolving the library symbols at run time. Third, it increases the security and stability of your program, as it prevents any malicious or incompatible changes to the libraries that your program uses.

添加您的观点

Sudeep K S

Firmware Engineer @ Intel Corporation | Vision IP Development | Camera Driver Development
举报内容
Advantages of static linked code are. Performance: Statically linked programs may perform slightly better than dynamically linked programs. This is because statically linked programs do not need to load any libraries at runtime. Portability: No external linkages required as all libs are statically linked to executable. Security: In some cases, security concerns may dictate that you statically link your program. This is because statically linked programs are less vulnerable to certain types of attacks. like DLL hijacking, code injection

已翻译

赞
William M.

Technology wrangler, autodidact, question asker.
举报内容
Choosing between dynamic linking versus static linking is similar to choosing between having problems now versus having problems later.

已翻译

赞

2 Drawbacks of static linking

Static linking also has some drawbacks that make it less desirable for other scenarios. First, it increases the size of your executable file, as it duplicates the library code for every program that uses it. This can lead to disk space and memory waste, as well as longer compilation and loading times. Second, it reduces the flexibility and maintainability of your program, as it prevents you from updating or replacing the libraries without recompiling your program. This can make it harder to fix bugs, add features, or support different platforms.

添加您的观点

Vikram vel

Firmware Engineer
举报内容
I've encountered several drawbacks of static linking in my programs on Linux. One major issue is increased binary size; static linking includes all library code in the executable, resulting in larger files that consume more memory. This can be particularly problematic on resource-constrained systems. Another drawback is reduced flexibility and security updates. With static linking, libraries aren't dynamically loaded, so security patches and updates require recompilation and redistribution of the entire program, making it cumbersome and potentially leaving vulnerabilities unaddressed for longer periods. In many scenarios, a combination of both static and dynamic linking is a more balanced approach, addressing these issues more effectively.

已翻译

赞

3 Benefits of dynamic linking

Dynamic linking has some benefits that make it preferable for most scenarios. First, it reduces the size of your executable file, as it shares the library code among multiple programs that use it. This can save disk space and memory, as well as speed up compilation and loading times. Second, it increases the flexibility and maintainability of your program, as it allows you to update or replace the libraries without recompiling your program. This can make it easier to fix bugs, add features, or support different platforms. Third, it enables the use of plugins or extensions that can extend the functionality of your program at run time.

添加您的观点

Christian Huaux

Senior Developer & Technical Architect | Skilled in High-Performance Computing, Game Engine Optimization, and Cross-Platform Development
举报内容
Modularity to only update one DLL is critical. The problem is that modularity needs to be different "per programmer" to be efficient. A low-level programmer will be "constantly" touching "math. h" or "types. h" or "base class," doing more rebuilding than a UX programmer who will appreciate having his module in a separate DLL. So having different build configurations (more than debug/release/master), not just per target, but per targeted users, is the way to go; Projects usually choose Dynamic and Static; It should be both to have the most significant impacts on your team's iteration time and comfort. And make the distribution, follow your goal rules for static and dynamic. Nobody says we have to follow those rules during development.

已翻译

赞
Sudeep K S

Firmware Engineer @ Intel Corporation | Vision IP Development | Camera Driver Development
举报内容
Program size: Dynamically linked programs are smaller than statically linked programs because they do not contain copies of all the library code that they need. Reduced memory usage: Dynamically linked programs use less memory than statically linked programs because they share libraries with other programs. Easier updates: If a library is updated, dynamically linked programs can simply use the updated library at runtime. Statically linked programs need to be recompiled with the updated library. Better compatibility: Dynamically linked programs are more compatible with different operating systems and distributions because they do not need to be linked with specific versions of libraries.

已翻译

赞

4 Drawbacks of dynamic linking

Dynamic linking also has some drawbacks that make it problematic for some scenarios. First, it complicates the deployment of your program, as you need to ensure that the libraries that your program depends on are available and compatible on the target system. Your executable file may not run if the libraries are missing, outdated, or corrupted. Second, it degrades the performance of your program, as it introduces the overhead of loading and resolving the library symbols at run time. Third, it decreases the security and stability of your program, as it exposes it to potential malicious or incompatible changes to the libraries that your program uses.

添加您的观点

Christian Huaux

Senior Developer & Technical Architect | Skilled in High-Performance Computing, Game Engine Optimization, and Cross-Platform Development
举报内容
Dynamic linking also has a significant impact on compilation time. as each library is individually linked. And it's a monothread task. On Windows, on a project with 40+ DLLs, 40+ mins for a rebuild all, Changed most of the DLLs to be static, rebuild all went down to 15 mins.

已翻译

赞

5 How to choose between static and dynamic linking

When it comes to deciding between static and dynamic linking for your C/C++ program on Linux, there is no one-size-fits-all answer. It all comes down to the trade-offs you are willing to make and the requirements you have. However, some general guidelines can help you make an informed decision. If you want a self-contained and portable executable file that can run on any Linux system without dependencies, optimize performance, and protect your program from external changes or interference, then static linking is the way to go. On the other hand, if you want a small and efficient executable file that can share library code with other programs, update or replace libraries without recompiling your program, or use plugins or extensions to enhance your program at run time, then dynamic linking is the better choice.

添加您的观点

Sudeep K S

Firmware Engineer @ Intel Corporation | Vision IP Development | Camera Driver Development
举报内容
This is the thumb rule I follow: If you are developing a library, you should generally create a shared library. This will allow other programs to link to your library dynamically. If you are developing a small, self-contained program, you may want to consider statically linking it. This will make your program more portable and easier to deploy. If you are developing a large, complex program, you should generally use dynamic linking. This will keep your program size small and make it easier to update.

已翻译

赞

6 How to perform static and dynamic linking

To perform static or dynamic linking for your C/C++ program on Linux, you need to use the appropriate compiler and linker flags and commands. For static linking, you need to specify the static library files (usually with the .a extension) that you want to link with your program, and use the -static flag to tell the linker to use them. For example, if you want to link your program with the math library, you can use: gcc -static -o myprog myprog.c /usr/lib/libm.a For dynamic linking, you need to specify the shared library files (usually with the .so extension) that you want to link with your program, and use the -L and -l flags to tell the linker where to find them and what to name them. For example, if you want to link your program with the math library, you can use: gcc -o myprog myprog.c -L/usr/lib -lm Note that you don't need to specify the full path or the extension of the shared library, as the linker will look for them in the standard locations and append the .so suffix. You can also use the -rpath flag to tell the linker where to search for the shared libraries at run time, if they are not in the standard locations. For more details and options, you can consult the manual pages of gcc and ld.

添加您的观点

7 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?

添加您的观点

Christian Huaux

Senior Developer & Technical Architect | Skilled in High-Performance Computing, Game Engine Optimization, and Cross-Platform Development
举报内容
Security, external DLL, also means DLL Injection. Size on the device has virtually no impact, as it's scarce to share DLL; programs now come with "their" own DLL most of the time. DLLs are also a pain on Windows when you need to update or compile as if the program is running, you can't write the DLL. Or update from the web. So be careful with locking yourself. Depending on what is previously installed on the OS where your program will be executed is a bad idea; you will spend more time—Debugging your client's OS than your code. Been there, done that. "Be your own Docker" (c) me 2023

已翻译

赞

Operating Systems

+ 关注

给文章评分

我们借助人工智能创建了此文章。您认为这篇文章怎么样？

很棒不太好

举报此文章

查看全部

How do you choose between static and dynamic linking for a C/C++ program on Linux?

1

2

3

4

5

6

7

1 Benefits of static linking

2 Drawbacks of static linking

3 Benefits of dynamic linking

4 Drawbacks of dynamic linking

5 How to choose between static and dynamic linking

6 How to perform static and dynamic linking

7 Here’s what else to consider

Operating Systems

给文章评分

感谢您的反馈

更多Operating Systems相关文章

更多相关阅读内容

How do you choose between static and dynamic linking for a C/C++ program on Linux?

1

2

3

4

5

6

7

1 Benefits of static linking

2 Drawbacks of static linking

3 Benefits of dynamic linking

4 Drawbacks of dynamic linking

5 How to choose between static and dynamic linking

6 How to perform static and dynamic linking

7 Here’s what else to consider

Operating Systems

给文章评分

感谢您的反馈

查看其他技能