Together, touch category bypass [Methodology of kernel learning] is sent, touch the category ---------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------One. The purpose of studying the kernel wants to do things, just right. To talk about the study of the core and exchange experience, the premise must have an appropriate and accurate purpose, and can harvest well. Many people have such a doubt, why should I study the kernel? We are not kernel developers (perhaps you will become Linus assistants in the future :-). This is poor, well known, Linux contains two levels of meaning, one is the kernel, the second is the release version, the former is the core essence, the latter is the application plan, both are not ruined. The frontier is slightly behind, thinking is not learning; the slightly, the latter is not thinking. Learning is a step-by-step, we learn Linux, first contact is the distribution version of each characteristic, easy and easy to get started, and gradually have a preliminary understanding of the Linux framework in the process of learning, it is to understand Linux world. Game rules. While your own ability is improved, you will find that the rate of progress seems to fall, why? That's because you open it is just your own memory instead of intelligence, the rules of the game are the core high-level packaging. It gives users and even programmers' interfaces are friendly (I believe many people are inhaused by UNIX), and this The interface has become very mature, and I just need to know. For example, the biggest difference between the release is the respective FHS (file directory framework) and package management mechanism, such as Gentoo's Portage (Port tree), and Emerge, as long as Debian's APT-GET, users as long as it is In this second, you can get a bundle in the world of this release, but the way is the same, that is, the identity rules. When you have been cooked for these rules, you will find that you are just from the user's angle, not from the developer's perspective, the difference is whether you have innovative in the Linux world. Compared to the Linux open source community at home and abroad, it is not difficult to find that our level is still very low, and many innovations are all occupied by foreign open source communities. We can only follow, in passive situation. Still illustrative, such as you are a LFS enthusiast, beginners act according to the LFS document, and the old birds, refer to LFS Hints to transform their own systems, but, I don't know, this is all non-nations. LFS documents are developed by foreigners, and LFS Hints is also written by foreigners. It is not that we should not introduce technology, but we see that we have lack innovation; such as UDEVs adopted in the new kernel (everyone should not be unfamiliar), it is just a logical I / O layer rule, but It is DEVFS to evolution, and everyone will learn with the open source community. Why can foreigners think about innovative UDEV? In the final analysis, that is, we don't pay attention to the most underlying research, based on the problem of high-level packaging, what is the use of Microsoft toy software? We always talk about our own release, but have you thought about it, not only putting ready-made kernels, but if this is the case, Redhat will not be the most popular version of the world. One (Redhat has a seat in most projects in the open source community, once a new version of the new version is exited by innovation technology).
It can be seen, based on the nuclear look, the purpose is 2: 1. Learn a serious power to grasp the innovation 2. I know, I know what it is, from the deep understanding of Linux, so that the daily operation of the release can touch bypass. On the second day of the study. Study the realm of the kernel -> practice -> The development of the country is now in the realm of the domestic nucleus lies in reading (including myself :-), and our final realm is development, which is the developer of the kernel. One, so that China's Linux business is independent! III. The attitude of the research of the kernel has been developed for ten years. The system becomes very large. It can be said that the root is confused. We read it. It is often difficult to directly master the overall situation. Combine our actual situation, it is best to use Linux from everyday. During the version, encounter some very wonderful problems, you should combine it to study the related items of the kernel, touch the class bypass. For example, when you play the problem of Kernel PANIC, we should find that there is such a sentence in the init / main.c, and then find the solution, and finally discover the necessary file system modules. It is necessary to compile the kernel (self-furniture contradiction), and, for example, we must transform the runtime script of the release version, you should notice that some scripts are running in kernel (such as LinuxRC), most scripts are running in the user (RCX.D), understanding the big direction, it is well encoded. Guerrilla Sanda, gradually surrounded by the large center, the core research has to have a good tool assistant, and there must be information (standing on the shoulders of the giant, seeing more), these I will use the top post. Together, you can touch. I hope everyone will communicate and make progress together. ============================================================================================================================================================================================================= =====================
Looking at this post, I turned a little violates the original interest of the post, and I thought I watched it later ...
