TOFFEEプロジェクトへようこそ ：： Linuxチャンネル

When you work on Network Namespaces (which is a feature Linux Kernel provides), you usually create bunch of virtual ethernet ports called as veth interfaces. The veth interface itself is a separate LK virtual network driver which offers this specific functionality. But sometimes besides veth you can also configure an optional interface called dummy interface. The dummy interface, just like veth (and other such virtual interfaces) is provided by LK via the driver drivers/net/dummy.c. The interesting aspect of dummy interface is that it serves as an alter ego of the loop-back localhost (as mentioned in the tldp.org article link below). Which means you can set any valid IP and this can serve as alternate local host ip other than 127.0.0.1. So to learn more, kindly watch my complete video series on this fascinating network interface called Dummy Interface.

Linux Kernel Tasklets and Work queues are somewhat similar to user-space process threads in terms of the functionality and of course not in terms of its operation or its internal architecture. You can schedule Tasklet(s) for a registered function to run later. So typically the Interrupt top half (of the interrupt handler) performs a small amount of critical work (not to be confused with critical section), and then schedules such as a tasklet to execute later at the bottom half.

When you work on Network Namespaces (which is a feature Linux Kernel provides), you usually create bunch of virtual ethernet ports called as veth interfaces. The veth interface itself is a separate LK virtual network driver which offers this specific functionality. But sometimes besides veth you can also configure an optional interface called dummy interface. The dummy interface, just like veth (and other such virtual interfaces) is provided by LK via the driver drivers/net/dummy.c. The interesting aspect of dummy interface is that it serves as an alter ego of the loop-back localhost (as mentioned in the tldp.org article link below). Which means you can set any valid IP and this can serve as alternate local host ip other than 127.0.0.1. So to learn more, kindly watch my complete video series on this fascinating network interface called Dummy Interface.

Linux Kernel Tasklets and Work queues are somewhat similar to user-space process threads in terms of the functionality and of course not in terms of its operation or its internal architecture. You can schedule Tasklet(s) for a registered function to run later. So typically the Interrupt top half (of the interrupt handler) performs a small amount of critical work (not to be confused with critical section), and then schedules such as a tasklet to execute later at the bottom half.

Learning Linux Kernel Programming is always fascinating and yet challenging. So generally you may tend to learn Kernel Module programming, since such a module can be dynamically plugged into running Linux Kernel. But this will lead to confusion, and many assume kernel source is mostly a collection of these modules. Which in reality is not. Not just that, when we say Kernel Module, its a vaguely defined term. The term Module (as we know) is nothing but a collection of APIs, bunch of variables and associated data-structures. Which may or may not be a plugable kernel module. If you ask me, I am a fan of wiring Linux Kernel Modules, which may not be necessarily a pluggable kernel module. It all boils down to the address space at which these modules function inside a monolithic Linux Kernel. Which is nothing but Linux Kernel's address space. Hence here is my detailed multi-episode Youtube video series on Linux Kernel modules, a big picture and the significance of the

Here is a my multi-episode Youtube video series of me doing a live code (Code with Kiran Series). Linux Kernel GPIO drivers of Raspberry Pi offers access/control directly via sysfs filesystem interface. But often people don't see this aspect, instead as a part of IoT, they assume these GPIO pins can be only accessed by some Python, Java or C/C++ code libraries. But in the real case those libraries are just an abstract layer written around the core Kernel GPIO driver's sysfs filesystem. So here is an attempt where we learn this aspect in systems architecture point of view. And then later try to attempt writing such user-space GPIO control code without any third-party libraries. Since the sysfs driver access is via regular files (user-space context), we can now use any language of our choice and access the same like any files. In that process we can attempt writing our own bunch of abstract APIs resembling a typical GPIO access library.

Here is a my multi-episode Youtube video series of me doing a live code (Code with Kiran Series). Linux Kernel source consists of many libraries within. For example such as to create and manage Linked Lists, String Operations, Encryption Algorithms, Compression Algorithms, and so on. In the Kernel, various Kernel Modules (including various Device Drivers) will use these Kernel inclusive Libraries. Unlike user-space programming, in Kernel Space when you code, you have to use these Libraries which are actually part of kernel (i.e sharing the same address space). And in some cases you can also make your own libraries and publish as well. In the case of user-space programs, we use third-party libraries (explicit or implicit), but that cannot happen when you code in kernel space. The libraries share the same address space and they are part of Linux Kernel mainline source. Hence it is important to understand the significance of using such library as demonstrated in my video series bel

Here is a code-walk of a git-hub source submitted by a viewer Sekhar Pariga, which is a Multi-client TCP Server with pthreads Mutex synchronized. Basically it is a Directory Listing Server, that serves ls, cd, pwd of directory functions to connecting TCP clients which are remotely (locally) executed by the server. Server is implemented using multi-thread pthread library, to serve multiple clients simultaneously. Server keeps the each accepted client session details in the queue data structure. Client session queue contains each clients socket-fd and present working directory as a queue node.

A detailed Youtube video series of various programming language performance and overheads - a big picture

When learning Linux Kernel programming, often I notice my students and viewers gets confused and they start with learning writing Linux Kernel modules. And so they develop the common misconception about Kernel Programming in general. They assume writing code in Linux Kernel means writing kernel modules. Which is absolutely not. Kernel modules are an optional choice and are part of Linux Kernel. But besides modules, Linux Kernel has lot of other mainstream code. Hence if anyone wants to be a Kernel Developer, you should be aware that sometimes you add new code via modules, sometimes without them. And if you ask me, I am not much in favor of writing Kernel modules. Instead in my code, I try to integrate and make them a part of Linux Kernel so that they all get initialized during boot time. Here is an extensive Youtube video of mine on Linux Kernel Programming, with and without Kernel Modules.

Learning Linux Kernel Programming is always fascinating and yet challenging. So generally you may tend to learn Kernel Module programming, since such a module can be dynamically plugged into running Linux Kernel. But this will lead to confusion, and many assume kernel source is mostly a collection of these modules. Which in reality is not. Not just that, when we say Kernel Module, its a vaguely defined term. The term Module (as we know) is nothing but a collection of APIs, bunch of variables and associated data-structures. Which may or may not be a plugable kernel module. If you ask me, I am a fan of wiring Linux Kernel Modules, which may not be necessarily a pluggable kernel module. It all boils down to the address space at which these modules function inside a monolithic Linux Kernel. Which is nothing but Linux Kernel's address space. Hence here is my detailed multi-episode Youtube video series on Linux Kernel modules, a big picture and the significance of the

The Linux Channel :: Weekly News Digest - Week 03 - June 2020 > Kimχ Micro: A powerful alternative to the Raspberry Pi that supports PCIe cards > Linux Lands And Then Reverts Usage Of Flexible Array Members > AMD Ryzen 4000-Powered Asus Mini PC Challenges Intel's NUC > Aaeon's Raspberry Pi-like Board Bears An Intel 8th Gen Core i7 CPU and 16GB of DDR4 > Linus Torvalds Announces Massive Linux Kernel 5.8 Update > New Cisco Cloud Scale ASIC & 400G Line Card Announced > Key Differences of PoE vs PoE+ vs PoE++ Switches a STH Guide