Chapter 1 Introduction

Transcription

1 Chapter 1 Introduction Hsung-Pin Chang Department of Computer Science National Chung Hsing University

2 Preference On the basis of of the Linux kernel Linux source code is contained in more than 8,000 C and assembly language files 4 million lines of code 144 megabytes of disk space

3 Resources for Tracing Linux Source code browser LXR (Source code navigator) Global Books Understanding the Linux Kernel, D. P. Bovet and M. Cesati, O'Reilly & Associates, Linux Core Kernel Commentary, In-Depth Code Annotation, S. Maxwell, Coriolis Open Press, The Linux Kernel, Version 0.8-3, D. A Rusling, Linux Kernel Internals, 2 nd edition, M. Beck et al., Addison-Wesley, Linux Kernel, R. Card et al., John Wiley & Sons, 1998.

4 Introduction Linux was initially developed by Linus Torvalds in 1991 based on Intel Open source under GNU Public License Commercial distributions collect vast additional software with the Linux to provide a fully operating system. The Linux source is usually installed in the /usr/src/linux directory

5 Kernel Source Code Organization

6 Linux Advantages over Commercial Competitors Linux is free Linux is fully customizable in all its components You are allowed to freely read and modify the source code of the kernel by the GPL (General Public License) Linux runs on low-end, cheap hardware platforms

7 Linux Advantages over Commercial Competitors (Cont.) Linux is powerful Fully exploit the hardware features And its main goal is efficiency Linux has a high standard for source code quality The Linux kernel can be very small and compact You may fit both a kernel image and full root file system on just one 1.4 MB floppy disk

8 Linux Advantages over Commercial Competitors (Cont.) Linux is highly compatible with many common operating systems You may mount file systems for all versions of operating system Linux support many network system, such as Ethernet, FDDI Linux can directly run programs written for other operating system by suitable libraries

9 Linux Advantages over Commercial Competitors (Cont.) Linux is well supported Very easy to get patches Numerous Device Driver support Numerous newsgroup and mailing list

10 Hardware Dependency Linux make a distinction between hardware-dependent and hardwareindependent source code Both arch and include directory include subdirectories that correspond to the hardware platform Alpha, arm, i386, ia64, m68k, mips

11 Linux Versions Linux distinguishes stable kernels from development kernels Linux version: X.Y.Z X.Y: Version number If the second number is even: stable kernel If the second number is odd: development kernel Z: release number Fix bugs reported by Users

12 Basic Operating System Concepts The operating system fulfill two main objectives Interact with the hardware components Provide an execution environment to applications Thus, when user wants a hardware resource Issue a request to the operating system

13 Basic Operating System Concepts (Cont.) Therefore, O.S. must rely on the availability of specific hardware features to forbid user applications to directly interact with hardware or memory CPU provides: Nonprivileged mode for user programs Privileged mode for the kernel Unix calls User Mode and Kernel Mode respectively

14 Multiuser Systems Multiuser O.S. includes several features An authentication mechanism for verifying the user s identify A protection mechanism against buggy user programs that could block other applications A protection mechanism against malicious user programs An accounting mechanism that limits the amount of resource units assigned to each user

15 Users and Groups All users are identified by a unique number called the User ID or UID If user want to use the computer Provide the Login name and password To selectively share material with other users Each user is a member of one or more groups Each group is identified by a unique number called a Group ID, or GID

16 Users and Groups Root is a special user that handle user accounts, perform maintenance task etc. Also called superuser, supervisor

17 Process Process An instance of a program in execution Execution context A process executes sequence of instructions in an address space The address space is the set of memory addresses that the process is allowed to reference Multiple sequence of instructions would be executed in the same address space, i.e., threading

18 Process (Cont.) Process/kernel model When a process makes a system call (i.e., a request to the kernel) The hardware change the privilege mode from User Mode to Kernel Mode Thus, the O.S. acts with the execution context of the process When system call complete, kernel forces the hardware to return to User Mode and the process continues its execution.

19 Kernel Architecture Module An object file whose code can be linked to the kernel at run time Advantages A modularized design approach Platform independence A disk driver for SCSI works at both IBM PC or HP Alpha Frugal main memory usage No performance penalty Once linked, module performs equivalent to the statically linked kernel

20 Unix Filesystem Overview Read it by yourself

21 An overview of Unix Kernels The Process/Kernel Model Process Implementation Reentrant Kernels Process Address Space Synchronization and Critical Regions Signals and Interprocess Communication Process Management Memory Mangement Device Drivers

22 The Process/Kernel Model Each CPU can run in either User Mode or Kernel Mode 80x86 has four different execution states Linux use only User Mode and Kernel Mode Each CPU also provides special instructions to switch between these modes

23 The Process/Kernel Model (Cont.) The process/kernel model assumes that processes that require a kernel service use specific programming construct called system calls The kernel itself is not a process but a process manager that implements many system calls

24 The Process/Kernel Model (Cont.) However, Unix systems also include a few privileged process called kernel threads They run in Kernel Mode and in kernel address space They do not interact with users They are usually created during system startup and remain alive until the system is shut down.

25 The Process/Kernel Model (Cont.) kernel routine can be activated in several ways A process invokes a system call The CPU executing the process signal an exception. A peripheral device issues an interrupt to the CPU Invoke interrupt handler A kernel thread is executed

26 Process Implementation Each process is represented by a process descriptor that PCB (Process Control Block) The kernel saves the contents of processor registers in the descriptor when performing context switch Program Counter (PC) and Stack Pointer (SP) General purpose registers Floating point registers Process control registers (Processor Status Word) containing information about the CPU state The memory management registers

27 Reentrant Kernels Unix kernels are reentrant Several processes may be executing in Kernel Mode at the same time On uniprocessor systems, only one process can progress, but many may be blocked in Kernel Mode waiting for some events

28 Reentrant Kernels (Cont.) Implementation approach Reentrant Functions: functions that only modify local variables but do not alter global data structures Kernel uses the nonreentrant functions and locking mechanism To ensure only one process can execute a reentrant kernels

29 Kernel Control Path A kernel control path denotes the sequence of instructions executed by the kernel to handle System call Exception Interrupt

30 Process Address Space Each process runs in its private address space Running in User Mode refers to private stack, data, and code areas When running in Kernel Mode, the process addresses the kernel data and code area and uses a private kernel stack Since several kernel control paths exists, thus each kernel control path uses its own private kernel stack

31 Process Address Space (Cont.) Memory sharing Several users uses the same program The code segment is shared by all users Processes may use shared memory as a kind of interprocess communication scheme mmap() Map file into a part of a process address space

32 Synchronization and Critical Regions Implement a reentrant kernel requires synchronization Different kernel control paths may access the same kernel data structure Called Race condition This section of code is called a critical region

33 Synchronization and Critical Regions (Cont.) Possible solution Atomic operation Read and decrement a variable with a single, noninterruptible operation However, many kernel data structures, e.g. linked list, cannot be accessed with a single operation

34 Synchronization and Critical Regions (Cont.) Solutions Nonpreemptive Kernels Interrupt disabling Spin locks

35 Nonpreemptive Kernels When a process executes in Kernel Mode, it cannot be arbitrarily suspended and substituted with another process Ineffective in multiprocessor system Two kernel control paths running on different CPUs may concurrently access the same data structure

36 Interrupt disabling Approach Disable interrupts Enter critical section Reenable interrupts Ineffective if the critical region is large Cannot work in a multiprocessor system

37 Semaphores Semaphore A counter Has two atomic operations Up Down Each semaphore is associated with a list of waiting process Link processes that are blocked Effective in both uniprocessor and multiprocessor

38 Spin Locks In Multiprocessor, semaphores may be inefficient if the time required to update the data structures is short Insert a process into a semaphore list is relative expensive Spin lock Executing a tight instruction loop until the lock becomes open Used in a multiprocessor system

39 Signals Signal: a mechanism to notify process of system events Asynchronous notification Synchronous errors or exceptions A process may react to a signal by Ignore the signal Asynchronously execute a specified procedure (the signal handler)

40 Signals (Cont.) A set of defined signals 1)SIGHUP 2) SIGINT 3) SIGQUIT 4) SIGILL 5) SIGTRAP 6) SIGIOT 7) SIGBUS 8) SIGFPE 9) SIGKILL 10) SIGUSR1 11) SIGSEGV 12) SIGUSR2 13) SIGPIPE 14) SIGALR 15)SIGTERM 17) SIGCHLD 18) SIGCONT 19) SIGSTOP 20) SIGTSTP 21) SIGTTIN 22) SIGTTOU 23) SIGURG 24) SIGXCPU 25) SIGXFSZ 26) SIGVTALRM 27) SIGPROF 28) SIGWINCH 29) SIGIO 30) SIGPWR

41 Signals (Cont.) If process does not specify its action, the kernel perform a default action Terminate the process Core dump and terminate the process Ignore the signal Suspend the process Resume the process s execution, if it was stopped

42 Interprocess Communication System V IPC Semaphores Message queues Shared memory

43 Process Management Create a process Fork(): create a new process Exec(): load a new program The process invokes a fork() is the parent, while the new one is its child

44 Process Management (Cont.) The original implementation of fork() duplicates both the parent s data and code and assign to the child Currently, Copy-On-Write approach _exit(): terminate a process Kernel release resources owned by the process Send a SIGCHLD signal to its parent process

45 Zombie Process The wait() system call allow a process to wait until one of its children terminates It returns the process ID (PID) of the terminated child Zombie process Terminated but before its parent executes wait() system call Still hold the task_struct data structure

46 Zombie Process (Cont.) The related data structure is released until wait() call But, how about a parent terminates without issue a wait() call? Start a child in background and then parent exits

47 Zombie Process (Cont.) The solutions rely on init process Created during system initialization When child terminated with no parent Change its parent to init Init then routinely issues wait() system call

48 Process Group Unix introduces the notion of process groups to represent a job $ ls sort more A progress group consists of three processes: ls, sort, more Login session All processes that are descendants of the login shell process

49 Memory Management Virtual memory

50 Random Access Memory RAM Usage A few megabytes are dedicated to storing kernel image (kernel code and kernel static data structures)

51 Random Access Memory Usage (Cont.) The remaining Satisfy kernel requests for buffers, descriptors, and other dynamic kernel data structures Satisfy process requests for generic memory area and for memory mapping of files Cache for hard disk

52 Kernel Memory Allocator Satisfy the requests for memory area from all parts of the system Both the kernel and user applications Features Fast Minimize the amount of wasted memory Reduce memory fragmentation problem Cooperate with other memory management subsystems

53 Process Virtual Address Space Handling A process s address space Contain all the virtual memory address that the process is allowed to reference Stored as a list of memory area descriptors

54 task_struct mm mm_struct count pgd mmap mmap_avl mmap_sem vm_area_struct vm_end vm_start vm_flags vm_inode vm_ops vm_next vm_area_struct vm_end vm_start vm_flags vm_inode vm_ops Process s Virtual Memory data code vm_next

55 Process Virtual Address Space Handling (Cont.) A process s virtual address space contains The executable code the program The initialized data of the program The uninitialized data of the program The initial program stack (i.e., the User Mode stack) The executable code and data of needed shared libraries The heap (for memory dynamically requested by program)

56 Process Virtual Address Space Handling (Cont.) Demand paging Loading virtual pages into memory as they are accessed Decide the page is in swap file or somewhere in disk

57 Swapping and Caching Swap area on disk Use physical memory as a cache Defer writing to disk Sync() system call force disk synchronization by writing all dirty pages All O.S. also periodically write dirty pages to disk

58 Device Drivers