Anatomy of the Linux Kernel Tim Jones Number of words 2730 Computer science content High Math content Low English language complexity Low Learning objectives + so acquire basic vocabulary related to ope + tounderstand Sub-areas covered + Linux kernel and tt eubsystems Keywords + emel - the central component of most computer operating systems (OS). Its functions include managing the system's resources (the communication between, hardware and software components) + Linux kemel - Unislike operating system kernel + VES(Virtual file system) -an abstraction layer on top of a move concrete filesystem + GNU-a computer operating system composed entirely of free software, initiated sm 1984 by Richard Stallman + GPL-a widely used free software license, criginally written by Richard Stallman for the GNU project + Minix- free/open source, Unix like operating system (OS) based on a miciokemel architecture + Unix a computer operating system originally developed in 1969 by a group of ATT ‘employees a Bell Labe including Ken Thompson, Dennis Ritchie and Douglas Ixoy + operating system - the software that manages the sharing of the resources of a computer and provides programmers with an interface used to access those + buffer-a region of memory used to temporatily hold data while tts being moved from one place to another + buffer cache - a col sored elsewhere or tion of data duplicating original values computed earlier, where the original data is expensive to fetch (owing t access time) or to compute, compared to the cost of reading the cache English++ 29English++ 30 Summary ‘As the title suggests, this aticle is about the linux kernel. I starts with a historical introduction, which icludes information on unix and minx the predecessors of linus Kernel in general and how itis constructed ‘The th ern 1d section is divided into cub sections which describe the subsystems in the lis J. Subsequent sections deal with memory management, process management divers layer or network stack. The article isnot very complex, being only an intsoduction, ‘tothe linux kemel It concludes with alist of inks to longer astcles about the linux kernel and its subsystems. Jak wskazuje tytul,artykul ten jest o jadrze linuxa, Na poczatku maly wstep histo- rycany. W te} czesci zawarte sq informacje o unix minixie, poprzednikach linwxa [Nastepna czgé¢ traktuje o jgdrze linuxa jako calosa, cay jak ogolnie jest ono zbudo wane, Tizeciacagé¢ jest slotona z mniejszych podczeéci. Kazda opisuje jeden 2 podsys temow jadra. Kolejne ezeéct mowig o: zarzgdzanta pamigcta, zarzadzantu procesamat vwarstwa sterownikow, systemic plikow, interfejsie wywolan systemowych oraz sto sie slecowym, Artykul nie jest zbyt rozbudowany, fest on oy nw do tnnycs axtykulow bardzle} vstepem do jadra Inuxa, Na kostcu tekstu znajduje sig lista odo: szczegélowo omawiaigcych roane cagéc jgdra Pre-reading exercises What ae the most popular operating systems? 2. What are the advantages of Linux? 3. What are the disadvantages of Linux?Anatomy of the Linux Kernel “The Limuxd kemel isthe core of a lage and complex operating system, and while i 4s huge, its well organized in terms of subsystems and layers In this article, you can explore the general structure ofthe Linux Kemel and get to know its major subsystems and core interfaces, Wher possible, you get links to other IBM articles to help you dig deeper. bis antic is to i ‘Given thatthe goal: duce you tothe Linux Kemel and explore ts architecture and major components, let's start with a short tour of Linux kernel history then lock at the Linux kernel architecture ftom 30,000 feet, and, finally examine ts _major subsystems, The Linux kemel is over six million lines of code, so this introduction is not exhaustive. Use the pointers to more content to dig in further. ‘A short tour of Linux history While Linux is arguably the most popular open source operating syst sm, its history 4s actualy quite short considering the timeline of operating systems, In the early days of computing. programmers developed on the bare hancware in the hardware’s Tanguage. The lack of an operating system meant that only one application (and one user) could use the laige and expensive device at a time. Eanly operating systems were developed in th 50s to provide a simpler development experience, Examples 1 System (GMOS) developed for the IBM 701 and the FORTRAN Monitor System (EMS) developed by North American Aviation for the sindlude the General Motors Oper 1M 709. In the 1960s, the Massachusetts Institute of Technology (M 1) and a host of companies developed an experimental operating system called Multics (or Multiplexed Information and Computing Service) for the GE-645, One of the developers ofthis operating system, ATAT. dropped out of Multics and developed their own operating system in 1970 called Unies. Along with this operating system was the C language, for which C was developed and then rewritten to make operating system development portable Andrew Tanenbaum created a microkemel version of UNIX®, NIX (for minimal UNIX, th ¢ ran on small personal computers. This open source operating system inspired Linus Torvalds’ initial development of Linux in the eaily 19905 Linux quickly evolved from a singleperson project to a worldwide development project involving thousands of developers. One of the most important decisions for [Linux was its adoption of the GNU General Public License (GPL). Under the GPL, the Linux kernel was protected from comm ial exploitation, and it also benefited from English++ operating system the softwave that manages the sharing of the resource: es ofa computer and provides programmers with an interface used to access those resources 31English++ buffer region of memory used to temporarily hold ata while itis being moved from one place to another ‘VES(Wirtual File Systers) an abstraction layer ‘on top of a more concrete 32. file system the userspace development of the GNU project (of Richard Stallman, whose source dwarfs that of the «2x Kernel). This allowed useful applications euch as the GNU Compiler Collection ( and various chell euppost Introduction to the Linux kernel Now on to a high-altitude look at the GNU/Linux operating system archit ‘an think about an operating system from two levels. ‘Av the top is the user, or application, space. This is where the user applications are ‘executed. Below the user space is the kernel space. Here, the Linux kemel exists There is also the GO ved nary (glibc). This provides the system call interface that ‘connects to the kemel and provides the mechanism to transition between the userspace application and the kemel, This is impoxtant because the kemel and user application ‘occupy different protected addzess spaces, And while each user space process occupies ts own virtual address space, the kemel occupies a si address space. For more ‘information, see the links in the resources section ‘The 1x kemel can be fusther divided into three gross levels, At the top is the system call interface, which implements the basi functions sch as ead and waite, Below the system call interface is the kemel code, which can be more accurately defined as the architecture independent kernel code, This code is common to all ofthe processor 3: chitectures supported by Linux Below thi is the architecture dependent code, which forms what is more commonly called a BSP (Board Support Package). This code serves as the processor and pl tform specific code for the piven architecture, Properties of the Linux kernel ‘When discussing the architecture of a large and complex system, you can view the system ym many perspectives. One goal of an archi tural decomposition is to vide a way to understand the source better and that’s what well do here, ‘The Linux kemel implements a number of important archi nal attibutes, Ata high level. and at lower levels, the kemel is layered into a number of distinct subsystems ‘Linux can also be considered monolithic because stlumps all ofthe basic services into the kemel. This fers from a microkesnel architecture, where the kemel provides basic services such as communication, 1/0, and memory and process management, and more specific services are plugged in to the microkemel layer. Bach has its own advantages, but Tl steer clear of that debate ‘Over time, the Linux kemel has become efficient in terms of both memory and CPU usage, as well as extremely stable, But the most interesting aspect f Linux, given itssize and complexity. sits portability Linux can be compiled to 1un on a huge number of procestors and platforms with different architectural constraints and needs, One ‘example is the ability of Linux to run on a process with a memory management unit (MMU}, as well as those that provide no MMU, The uClinux port of the Linux kemel provides for non-MMU support See the resources section for more details Major subsystems of the Linux kernel [Now let's look at some of the major components of the Linux kemel using the breakdown, System call interface The Sc1 a thin layer that provides the means to perform function calls from user epace into the Kemel, As discussed previously, this interface can be architecture dependent, even within the same processor family. The SCI is actually an interesting function-call multiplexing and demultiplexing service, You can find the SCI implementation in flinuxy/kemel, as well as architecturedependent portions in Ainux(auch, More details for this component are available tn the Process management Process management ie focused on the execution of processes. In the kemnel, these are called threads and represent an individual vistualization of the processor (thread code, data, stack, and CPU registers). Im user space, the term process is typically used ‘though the Linux implementation does not separate the two concepts (processes and interface (API) Uhrough the SCI $k, exec, or Portable Operating System Interface [POSTX) threads). The kemel provides an application program to create a new process ff functions), stop a process (hill, exit), and communicate and synchronize between ‘them (signal, or POSIX mechanisms) ‘Also in process management there is a need to share the CPU between the active ‘threads, The kernel implements a novel scheduling algorithm that operates in constant 6 O12) the same amount of time is taken to schedule one thread as, ‘me, regardless of the number of threads vying for the CPU. This is calle it isto schedule many. The O(2) scheduler also supports multiple processors (called Symmetric MultiProcessing, or SMP). You can find the process management sources in flinux/kemnel and architecturedependent sources in J/linux/arch). You can learn, ‘more about this algorithm in the resources section, Memory management Another important resource that's managed by the kemel is memory For efficiency aiven the way that the hardware manages virtual memory, memory is managed in English++ Linux kernel Unixlike operating system kernel kemel the central component of most computer operating systems (05) Its factions include managing the system's (the communication between hardware and software components 33English++ freejopen source, Unbelike operating based ona system (05) ‘microkernel auchitecture ort. widely used free software license, coviginally wnitten by Richard Stallman for the 34 GNU project what are called pages (4KB in size for most architectures). Linux includes the means to manage the available memory, as well as the haudware mechanisms for physical and vistual mappings. ‘But memory management is much more than. naging 4KB buffers, Linux provides abstractions over 4KB buffers, such as the slab allocator. This memory management scheme uses 4K2 buffers as its base. but then allocates structures from within, keeping ‘tack of which pages are full, partially used, and empty. This allows the scheme to dynamically grow and shrink based nn the needs of the greater system Supporting multiple users of memory there ate times when the available memory can be exhausted, For this reason, pages can be moved out f memory and onto the disk This process is called swapping because the pages ate swapped ftom memory conto the hard disk. You can find the memory management sources in /inux/mm. Virtual file system The virtual file aystem (VPS) és an interesting aspect of the Linux kernel because provides a common interface abstraction for file systems. The VFS provides a switching layer between the SCI and the file systems supported by the kernel. {At the top of the VES is a common API abstraction of functions such as open, close, read and write. At the bottom of the VFS ate the file system abstractions that define how the upper layer functions are amplemented, These are plugins for the given file system (of which over 50 exist). You can find the file eystem sources in /linufs Below the filesystem layer is the buff to the filesystem layer independent of any particular file eystem). This caching layer ashi cache, which provides a common set of fun suons ‘optimizes access to the physical devices by keeping data around ime (or speculatively read ahead so that the data is available when needed). Below the buffer ‘cache are the device drivers, which implement the intesface forthe particular physical device Network stack ‘The network stack, by design, follows alayexed arc {ecture modeled after the protocols themselves. Recall thatthe Internet Protocol (1) is the core network layer protocol that sits below the transport protocol (most commonly the Transmission Contiol Protocol ‘or TCP}, Above TCP isthe sockets layer, which is invoked through the SCL ‘The sockets layer is the standard API to the networking subsystem and provides a user interface to a vatiety of networking protocols. From raw frame access to IP protocol data units (PDUs) and up to TCP and the User Datagram Protocol (UDP), the socketslayer provides a standardized way to manage connections and move data between ‘endpoints. You can find the networking sources in the kemel at imuwnet Device drivers ‘The vast majority of the source code in the Linux kernel exists im device drivers that rake a particular hardware device usable. The Linux source tree provides a drivers subdirectory that is further divided by the various devices that are supported, such as ‘Bluetooth, 12C, serial and so on. You can find the device driver sources in /linux/chivers, Architecture-dependent code Whale much o sax is Independent of the architecture om which it runs, there ate elements that must consider the architecture for normal operation and for efficiency Te inux/arch subdirectory defines the architecturedependent portion of the kemel source contained in a number of subdirectories that are specific to the architecture (collectively forming the BSP). For a typical desktop, the 1386 directory is used. Each architecture subdixectory contains a number of other subdiectories that focus on a particular aspect ofthe kernel, such as boot kemel, memory management, and others ‘You can find the architecture-dependent code in /linux/atch, Interesting features of the Linux kernel If the portability and efficiency ofthe Linux kemel weren't enough it provides some other features that could not be classified in the previous decomposition. Linux. being a production operating system and open source, is a great test bed for new protocols and advancements of those protocols. Linux supports a large number of networking protocols. including the typical TCP/IP and also extension for high-speed networking (greater than 1 Gigabit Bthemet (GbE] and 10 GbE). Linux also supports protocols such as the Stream Control Transmission Protocol (SCTP), which provides many advanced features above TC? (as a replacement transport level protocol) Linux is also a dynamic kernel, supporting the addition and removal of software components on the fly: These are called dynamically loadable kemel modules, and they can be inserted at bot ‘hen they're needed (when a particular device is found requiring the module) or at any time by the user A recent advancement of Linux is its use as an operating system for other operating systems (called a hypervisor). Recently, a modification to the kemel was made called the Kernel based Virtual Machine (KVM). This modification enabled a new interface to user space that allows other operating systems to run above the KVM-enabled soft® Windows® can kernel, fn addition to running another instance of Laux, Mi English++ env a computer operating system composed entizely of free software initiated in 1984 by Richard Stallman 35English++ buffer cache collection of data duplicating ‘original values stored. elsewhere or computed ‘earlier, where the original data is expensive to fetch (owing to longer access time) or to compute, compared to the cost of reading the cache 36 also be virtualized. The only constraint is that the underlying processor must suppoxt the new vi jon instructions, See the resouuce section for more information, Going further ‘This article jut scratched the eusface of the Linux kernel architecture and its features and capabilities. You can check out the Documentation directory that is provided in ‘every Linux distribution for detailed information about the contents of the kernel Resources: + The GNU site (http./www-gnu org/licenses) describes the GNU GPL that covers the Linux kernel and most useful applications provided with st, so described is less restrictive form of the GPL « the Lesser GPI. (LGPL) su NIX (http:/fen wikipedia orghviki/Unics), MINIX.(http://en.wikipedia org/wiki/ Minix) and Linux (http.//en wikipedia org/wiki/Linux) are covered in Wikipedia along with a detailed family tree of the operating systems, + Theat of the standard Cl UC Library (httpy/www gnu oxg/software/Mbch, o gb. is the implementation ary It's used in the GNU/Linux operating system, as wellas the GNU/Hlurd (http: /dixectory fof oxg/bued html) microkesnel operating system. + uclinux (hitpy/tewwwuclimuxorgh is a port of the Linux kemel that can execute fon systems that lack an MMU. This allows the Linux kernel to run on very small ‘embedded platforms, such as the Motorola DragonBall processor used in the PalmPot Personal Digital Assistants (PDA) + “Keel command using Linux system calls” (attps/wwwibm.comvdeveloper: works/linux/ibrary/Lsystem-alls/) (developerWorks, March 2007) covers the SCI which is an important layer in the Linux kemel, wit userspace support from alibe that enables function calls between user space and the kernel + “Inside the Linux scheduler” (hitp,/hvww ibm com/developerworkslinux/baryt- scheduler) (developerWorks, June 2006) explores the new O() scheduler introduced fn Linux 2.6 that is efficent, scales with a large number of processes (threads) and takes advantage of SMP systems. + “Access the Linux kernel using the /proc filesystem" (http//wwwibm.com/devel operworks/linuxtbrary/l-proc html) (developerWorks, March 2006) looks at the ‘proc filesystem, which isa virtual file system that provides a novel way for user space applications to communicate with the kemel. This article demonstrates / oc, as well as loadable kernel modules. + "Server clinic: Put vistual filesystems to work" (http //www ibm. comvdevelopes: wworks/linulibrary/Lscl2 html) (developer Works. April 2003) delves into the VES layer that allows Linux to support a vattety of different common interface. This same interface is also used for of such as socketside the Linux boot process’ (httpy/wnwwibmcom/developerworks/inwlibraty/ imuxboov index html (developer Works May2006) examinestheLinuxbootprocess, which taker cate of bringing up a Linux eystem and is the same basic process whether you'te booting from a hard disk, floppy. USB memory stick, of over the network, “Linux inital RAM disk (init) overview" (http wwwribm.com/developerworks/ linua/ibrary/linited html) (developes Works, July 2006) inspects the intial RAM disk Which isolates the boot process from the phystal medium from which i's booting Better networking with SCTP" (httpy/hvwwibm com/developerworks/linu/ library/-sctp) (developerWorks, February 2006) covers one of the most interest o networking protocols, Stream Control Transmission Protocol, which operates like TCP but adds a number of useful features such as messaging, multshoming and multsetveaming. Linux, like BSD, is a great operating eystem if you're interested sm networking protocols. “AnatomyoftheLinuxslaballocator” (http. //www som. com/developerworks/linux/ Ubrary/ Linux slaballocatoy) (developerWorks, May 2007) covers one of the most interesting aspects of memory management in Linus, the slab allocator. This mechanism originated in SunOS, but i's found a stendly home inside the Linux. kernel. ‘Virtual 1 com/developerworks/tinux/library/linuxvitt) ax" (http rw (developerWorks, December 2006) shows how Linux can take advantage of processors with vitualization capabilities Tanux and symmetric multiprocessing” {http /wvww thm com/developerworks/ library/linux-smp/) developerWorks, March 2007) discusses how Linux can also take advantage of processors that offer chiplevel multiprocessing ~DiscovertheLinuxKemelVirtualMachine" (http //wrww sbm. conv developerworks/ linuvlibraryllimwckwn) (evelopes Works, Apuil 2007) covers the recent introduction of virtualization into the kernel, which turns the Linux kes snto.a hypervisor for other virtualized operating systems, Checkout Tim's book GNU/Linux Application Programming httpy/wrww charlesriver comyBooks/BockDetailaspx?productID =91525) for more information on programming Lapux in user space In the developerWorks Linux zone { tp:/hwww ibm. conv/developerworks/linux), findmoreresourcesforLinuxdevelopers,including!smuxtutorials http:/www:sbm, com/developerworksiviews/linuw/ibraryvieweptype_by=Tutorials) aswell as our readers’ favorite Linux atticles and tutorials (nttp:wrwibm.com/developerworks/ linua/ibrary/-top-10 html] over the last month Stay curent with developerWorks technical events and Webcasts (http:/wwwibm, 1O5AGXO3&S_CMP=art) com/developerworks/offers/techbriefings/"$_TACI English++ Unix a computer operating system oxiginally developed in 1969 by a group of ATT employees a Bell Labs Ancluding Ken Thompson, Dennis Ritchie and Douglas Thoy 37