Presentation is loading. Please wait.

Presentation is loading. Please wait.

제2장. 파일 입출력 (File I/O) © 숙대 창병모.

Similar presentations


Presentation on theme: "제2장. 파일 입출력 (File I/O) © 숙대 창병모."— Presentation transcript:

1 제2장. 파일 입출력 (File I/O) © 숙대 창병모

2 목표 파일의 구조 및 특성을 이해한다. 파일을 열고 닫는다. 파일로부터 데이터를 읽고 쓴다. 파일 현재 위치 변경
기타 파일 제어 © 숙대 창병모

3 2.1 파일 구조 © 숙대 창병모

4 What is a file ? a file is a contiguous sequence of bytes
no format imposed by the operating system each byte is individually addressable in a disk file a file is also a uniform interface to external devices © 숙대 창병모

5 File Descriptor File descriptor open() returns a fd, an integer value
열린 파일을 나타내는 번호 used in subsequent I/O operations on that file close(fd) closes that file described by fd all of a process's open files are automatically closed when it terminates © 숙대 창병모

6 File Descriptor file descriptor : 0 ~ 19 © 숙대 창병모

7 User and Kernel Mode Kernel User process open(char *name, int mode) {
Address of kernel close() result=open(“/usr/glass/file.txt” , O_RDONLY); Address of kernel open() User code Address of kernel write() open(char *name, int mode) { <Place parameters in registers> <Execute trap instruction, switching to kernel code > <Return result of system call> } kernel code for open() { <Manipulate kernel data> . . . <Return to user code> } Kernel system call code C runtime library © 숙대 창병모

8 Inode (Index node) 한 파일은 하나의 i-node를 갖는다. 파일에 대한 모든 정보를 가지고 있음
file type : regular, directory, block special, character special, etc file size file permissions the owner and group ids the last modification and last access times if it's a regular or directory, the location of data blocks if it's a special file, device numbers if it's a symbolic link, the value of the symbolic link © 숙대 창병모

9 파일을 위한 커널 자료 구조 fd= open(“file”, O_RDONLY); 1 2 3 Inode list
Open file table Active inode table Fd table (per process) Disk Inode list 1 2 File pos refcnt=1 ... 3 File access ... File size File type Data blocks © 숙대 창병모

10 Process table entry 프로세스 테이블 (Process table)
커널(kernel) 자료구조 프로세스 목록 프로세스  프로세스 테이블 항목 프로세스 테이블 항목 (Process table entry) 파일 디스크립터 배열(file descriptor array) 포함 fd array © 숙대 창병모

11 Open File Table 파일 테이블 (file table) 파일 테이블 항목 (file table entry)
커널 자료구조 열려진 모든 파일 목록 열려진 파일  파일 테이블의 항목 파일 테이블 항목 (file table entry) 파일 상태 플래그 (read, write, append, sync, nonblocking,…) 파일의 현재 위치 (current file offset) i-node에 대한 포인터 © 숙대 창병모

12 Active i-node table Active i-node table i-node i-node table vs. i-node
커널 내의 자료 구조 Open 된 파일들의 i-node를 저장하는 테이블 i-node 하드 디스크에 저장되어 있는 파일에 대한 자료구조 한 파일에 하나의 i-node 하나의 파일에 대한 정보 저장 소유자, 크기 파일이 위치한 장치 파일 내용 디스크 블럭에 대한 포인터 i-node table vs. i-node © 숙대 창병모

13 2.2 파일 관련 시스템 호출 open() - 열기 creat() - 파일 생성 close() - 닫기 read() - 읽기
write() - 쓰기 lseek() - 이동 © 숙대 창병모

14 open() - 열기 #include <sys/types.h> #include <sys/stat.h>
#include <fcntl.h> int open (const char *pathname, int oflag, [ mode_t mode ]); 파일을 연다 파일이 없으면 경우에 따라 새로 만들어 질 수도 있다 리턴 값 : 파일 디스크립터(file descriptor), 실패하면 -1 pathname : 파일의 이름 mode : 파일의 access permission 값. 생략가능. 새로운 파일을 만드는 경우에만 사용됨 creat() 함수 설명 참조 © 숙대 창병모

15 open() 의 파라미터 두 번째 파라미터 oflag는 다음 상수들의 OR 이다 반드시 하나 지정해주어야 할 값
예 int fd; fd = open("afile", O_RDWR | O_CREAT, 0600 ); 반드시 하나 지정해주어야 할 값 O_RDONLY : 읽기 모드, write 함수를 사용 할 수 없음 O_WRONLY : 쓰기 모드, read 함수를 사용 할 수 없음 O_RDWR : 읽고 쓰기 모드, read write 사용 가능 © 숙대 창병모

16 open() 의 파라미터 선택적 지정 가능 O_APPEND: 모든 write 된 데이터는 파일의 끝에 추가
O_CREAT : 파일이 없는 경우 파일 생성 세 번째 인자 mode O_EXCL : O_CREAT이고 그 파일이 이미 있으면 에러 O_TRUNC : 파일이 있는 경우 파일 크기를 0으로 만든다 O_NONBLOCK : nonblocking 모드로 입출력을 함 O_SYNC : 각 write 함수 호출은 디스크에 물리적으로 쓰여진 후 리턴한다 © 숙대 창병모

17 예제: open.c /* open.c */ #include <stdlib.h>
#include <fcntl.h> #include <unistd.h> int main() { int fd; char fname[] = "afile"; if ((fd = open (fname, O_RDWR)) == -1) perror(fname); printf("%s is opened.\n", fname); close(fd); return 0; } © 숙대 창병모

18 creat () - 파일 생성 #include <sys/types.h>
#include <sys/stat.h> #include <fcntl.h> int creat ( const char *pathname, mode_t mode ); 새로운 파일을 생성한다 리턴 값 : 파일 디스크립터, 실패하면 -1 pathname : 생성하고자 하는 파일의 이름 mode : 파일의 access permission 값 © 숙대 창병모

19 creat () 의 파라미터 두 함수 호출은 동일 두 번째 인자 mode는 permission mode
fd = creat ( pathname, mode ); fd = open ( pathname, O_WRONLY | O_CREAT | O_TRUNC, mode); 두 번째 인자 mode는 permission mode 0644 0777 0444 © 숙대 창병모

20 close () - 닫기 #include <unistd.h> int close ( int fd );
작업이 끝난 후 파일을 닫는다. 리턴 값 : 성공하면 0, 실패하면 -1 fd : 닫고자 하는 파일의 파일 디스크립터 프로세스가 종료되면 모든 열려진 화일들은 자동적으로 닫힌다 © 숙대 창병모

21 블록 입출력 이진 파일 블록 단위 입출력 예 모든 데이터를 바이트의 연속으로 저장 이미지 파일, 데이터 파일, 실행 파일 등
한번에 일정한 크기의 데이터(블록)를 읽거나 쓰는 것 한 번에 하나의 레코드를 읽거나 쓰는 것 © 숙대 창병모

22 read () - 읽기 fd가 나타내는 파일에서 데이터를 읽는다 리턴 값: buf nbytes
#include <unistd.h> ssize_t read ( int fd, void *buf, size_t nbytes ); fd가 나타내는 파일에서 데이터를 읽는다 리턴 값: 성공하면 읽은 바이트 수 파일의 끝을 만나면 0 실패하면 -1 buf 읽은 데이터를 저장할 메모리의 시작 주소 nbytes 읽을 데이터의 바이트 수 © 숙대 창병모

23 read () - 읽기 읽을 데이터가 충분하면 한 번에 nbytes 만큼 읽는다.
파일의 끝에서 네트웍 입출력에서 size_t : unsigned integer ssize_t : signed integer © 숙대 창병모

24 예제 /* count.c: 파일의 문자 수를 센다*/
#include <stdlib.h> #include <fcntl.h> #include <unistd.h> #define BUFSIZE 512 int main() { char buffer[BUFSIZE]; int fd; ssize_t nread; long total = 0; if ((fd = open("afile", O_RDONLY)) == -1) perror("afile"); /* 파일의 끝에 도달할 때까지 반복 */ while( (nread = read(fd, buffer, BUFSIZE)) > 0) total += nread; close(fd); printf ("total chars in afile: %ld\n", total); return 0; } © 숙대 창병모

25 write() - 쓰기 fd가 나타내는 파일에 데이터를 쓴다 리턴 값 buf nbytes
#include <unistd.h> ssize_t write (int fd, void *buf, size_t nbytes); fd가 나타내는 파일에 데이터를 쓴다 리턴 값 성공하면, 파일에 쓰여진 데이터의 바이트 수 실패하면, -1 buf 쓸 데이터를 저장하고 있는 메모리의 시작주소 nbytes 쓸 데이터의 바이트의 수 © 숙대 창병모

26 예: Copy Files #include <stdio.h> #include <fcntl.h>
main(argc, argv) int argc; char *argv[ ]; { int fdin, fdout, n; char buf[BUFSIZ]; if (argc != 3) { fprintf(stderr,"Usage: %s filein fileout\n",argv[0]); exit(1); } if ((fdin = open(argv[1], O_RDONLY)) == -1) { perror(argv[1]); exit(2); } if ((fdout=open(argv[2],O_WRONLY|O_CREAT|O_TRUNC, 0644))==-1){ perror(argv[2]); exit(3);} while ((n = read(fdin, buf, BUFSIZ)) > 0) write(fdout, buf, n); exit(0); } © 숙대 창병모

27 임의 접근 순차 접근(sequential access) 임의 접근(random access)
파일의 처음부터 끝까지 순차적으로 진행하면서 데이터를 쓰거나 읽음 임의 접근(random access) 파일 내의 위치에 관계없이 원하는 자료가 있는 곳을 직접 접근 seek() 함수를 사용하여 입출력을 할 위치로 바로 이동 © 숙대 창병모

28 파일 위치 포인터 현재 파일 위치(current file position)
파일 위치 포인터(file position pointer) © 숙대 창병모

29 lseek() - 이동 #include <sys/types.h> #include <unistd.h>
off_t lseek (int fd, off_t offset, int whence ); 파일의 현재 위치(current file offset)를 이동 성공하면 현재 위치를 리턴, 실패하면 -1 whence : 위치 기준점 SEEK_SET : 파일의 시작점을 기준으로 이동 SEEK_CUR : 현재 위치를 기준으로 이동 SEEK_END : 파일의 끝을 기준으로 이동 offset : 기준점에서의 상대적인 거리 (byte 단위) SEEK_CUR, SEEK_END 와 같이 쓰일 때는 음수도 가능 © 숙대 창병모

30 lseek() – 이동 파일의 현재 위치(current file offset) lseek()
파일을 처음 열면 현재 위치는 0 즉 파일의 시작 파일에 대한 읽기/쓰기는 파일의 현재 위치에서 실행된다 읽기/쓰기 후 파일의 현재 위치는 읽기/쓰기 한 byte 수만큼 자동적으로 뒤로 이동 lseek() 임의의 위치로 파일의 현재 위치를 이동할 수 있다 © 숙대 창병모

31 예 :lseek rewind append record position increase file lseek(fd, 0L, 0);
loc = lseek(fd, 0L, 1); increase file lseek(fd, (long) MAX*sizeof(record), 2); write(fd, (char *) &reocrd, sizeof(record)); © 숙대 창병모

32 예제: /* lseek1.c */ #include <unistd.h> /* lseek1.c */
#include <fcntl.h> char buf1[] = "abcdefghij"; char buf2[] = "ABCDEFGHIJ"; int main() { int fd; if ( (fd = creat("file.hole", 0644)) < 0) perror("file.hole"); if (write(fd, buf1, 10) != 10) /* offset now = 10 */ perror("buf1"); if (lseek(fd, 40, SEEK_SET) == -1) /* offset now = 40 */ perror("lseek"); if (write(fd, buf2, 10) != 10) /* offset now = 50 */ perror("buf2"); return 0; } © 숙대 창병모

33 예제: /* lseek1.c */ lseek1.c 의 출력 파일 file.hole의 내용 1 2 3 4 5 6 7 8 9 a
1 2 3 4 5 6 7 8 9 a b c d e f g h i j 10 \0 20 30 40 A B C D E F G H I J © 숙대 창병모

34 예제: /* lseek2.c */ #include <sys/types.h> /* lseek2.c */
#include <unistd.h> #include <fcntl.h> int main() { int fd; off_t fsize; char buf[11]; if((fd=open("file.hole", O_RDONLY)) < 0) perror("file.hole"); fsize = lseek(fd, 0, SEEK_END); printf("size: %lu\n", fsize); lseek(fd, 40, SEEK_SET); read(fd, buf, 10); buf[10] = ‘\0’; puts(buf); return 0; } © 숙대 창병모

35 레코드 읽기 read(fd, (char *) &record, sizeof(record)); record READ
current file offset READ © 숙대 창병모

36 레코드 쓰기 write(fd, (char *) &record, sizeof(record)); record WRITE
current file offset WRITE © 숙대 창병모

37 레코드 수정 struct XXX record; …
read(fd, (char *) &record, sizeof(record)); … update record … lseek(fd, (long) -sizeof(record), 1); write(fd, (char *) &record, sizeof(record)); © 숙대 창병모

38 레코드 수정 record READ current file offset © 숙대 창병모

39 레코드 수정 record LSEEK current file offset © 숙대 창병모

40 레코드 수정 record WRITE current file offset © 숙대 창병모

41 2.3 시스템 호출 구현 © 숙대 창병모

42 Block I/O I/O is always done in terms of blocks
Sequence of a read() system call © 숙대 창병모

43 A Typical Hardware System
main memory I/O bridge bus interface ALU register file CPU chip system bus memory bus disk controller graphics adapter USB mouse keyboard monitor I/O bus Expansion slots for other devices such as network adapters. © 숙대 창병모

44 Reading a Disk Sector: Step 1
CPU chip CPU initiates a disk read by writing a command, logical block number, and destination memory address to a port (address) associated with disk controller. register file ALU main memory bus interface I/O bus USB controller graphics adapter disk controller mouse keyboard monitor disk © 숙대 창병모

45 Reading a Disk Sector: Step 2
main memory ALU register file CPU chip disk controller graphics adapter USB mouse keyboard monitor I/O bus bus interface Disk controller reads the sector and performs a direct memory access (DMA) transfer into main memory. © 숙대 창병모

46 Reading a Disk Sector: Step 3
main memory ALU register file CPU chip disk controller graphics adapter USB mouse keyboard monitor I/O bus bus interface When the DMA transfer completes, the disk controller notifies the CPU with an interrupt (i.e., asserts a special “interrupt” pin on the CPU) © 숙대 창병모

47 I/O 효율 디스크는 메모리에 비해서 매우 느린 장치 성능을 위해서 디스크 I/O의 횟수 최소화가 바람직
write1.c 에서 BUFFSIZE 값을 변경하며 수행 시간을 측정 8192 까지는 BUFFSIZE 가 클수록 성능이 향상 8192 보다 큰 경우는 성능에 변화가 없다 © 숙대 창병모

48 I/O 효율 BSD fast file system 의 I/O 단위는 8192 바이트 BUFSIZE가 8192 보다 큰 경우
성능이 하락할 수 있다 © 숙대 창병모

49 예제: /* write1.c */ #include <sys/types.h> /* write1.c */
#include <sys/stat.h> #include <fcntl.h> #define BUFFSIZE 512 #define FILESIZE (100 * 1024 * 1024) #define COUNT FILESIZE / BUFFSIZE int main() { int i, fd; char buf[BUFFSIZE]; memset(buf, '.', BUFFSIZE); if ((fd = creat("file.write", 0600)) < 0) perror("file.write"); for (i=0; i < COUNT; ++i) write(fd, buf, BUFFSIZE); close(fd); return 0; } © 숙대 창병모

50 read/write 구현 read/write 시스템 호출 파일 크기 증가 파일을 O_APPEND로 열면
file table entry의 file position 값이 읽은/쓰여진 바이트 수 만큼 증가 파일 크기 증가 Write 후 file position 이 i-node의 file size보다 크면 파일 크기 증가 파일을 O_APPEND로 열면 먼저 file position 을 i-node 의 file size 값으로 설정한 후 write 수행 따라서 모든 write는 파일의 끝에 덧붙여진다. © 숙대 창병모

51 read/write 시스템 호출 구현 1 2 3 Inode list Data blocks Open file table
Active inode table Fd table (per process) Disk Inode list 1 2 File pos refcnt=1 ... 3 File access ... File size File type Data blocks © 숙대 창병모

52 Example read(fd, buf1, 100); read(fd, buf2, 200);
copy the first block from disk into an buffer copy the first 100 bytes from the buffer into buf1 offset <- 100 read(fd, buf2, 200); copy the next 200 bytes from the buffer into buf2 read(fd, buf3, 5000); copy the remaning (3796bytes) from the buffer to buf3 copy the second block from disk into an buffer copy the remainig(1204 bytes) from the buffer into buf3 write(fd, buf4, 100); write(fd, buf5, 4000); © 숙대 창병모

53 lseek 구현 lseek 시스템 호출 SEEK_END를 사용하여 파일의 끝으로 이동하면
file table entry 의 current file offset 값을 변경 disk I/O 는 없음 SEEK_END를 사용하여 파일의 끝으로 이동하면 file table entry 의 current file offset 값은 i-node 의 파일 크기 값으로 설정됨 © 숙대 창병모

54 lseek 구현 lseek(fd, 300, L_SET) 1 2 3 Open file table
Active inode table Fd table (per process) Disk Inode list 1 2 File pos=300 refcnt=1 ... 3 File access ... File size File type Data blocks © 숙대 창병모

55 2.4 파일 공유 © 숙대 창병모

56 dup(), dup2() 사용 중인 파일 디스크립터의 복사본을 만듦 리턴 값 #include <unistd.h>
int dup (int fd); int dup2 (int fd, int fd2); 사용 중인 파일 디스크립터의 복사본을 만듦 dup()는 새 파일 디스크립터 번호가 할당됨 dup2()는 fd2 를 사용 리턴 값 성공하면 복사된 새 파일 디스크립터, 실패하면 -1 dup() 함수는 할당 가능한 가장 작은 번호를 리턴한다. dup2() 함수는 fd2 를 리턴한다. © 숙대 창병모

57 dup 구현 fd2 = dup(3); 1 2 34 Open file table Active inode table
Fd table (per process) Disk Inode list 1 2 File pos=300 refcnt=2 ... 34 File access ... File size File type Data blocks © 숙대 창병모

58 dup 구현 fd2 = dup(3); 1 2 34 Open file table Active inode table
Fd table (per process) Disk Inode list 1 2 File pos=300 refcnt=2 ... 34 File access ... File size File type Data blocks © 숙대 창병모

59 예제 /* dup.c */ #include <fcntl.h> #include <unistd.h>
int main() { int fd1, fd2, n; char buf[11]; if ((fd1 = open("afile", O_RDONLY)) < 0) perror("afile"); if ((fd2 = dup(fd1)) < 0) perror("dup"); n = read(fd1, buf, 10); buf[n] = ‘\0’; puts(buf); n = read(fd2, buf, 10); close(fd1); close(fd2); return 0; } © 숙대 창병모

60 dup2() newfd = dup2(3,1); 호출 후
리턴 값은 1 file table의 file position과 fd status flags는 공유됨 file descriptors의 fd flags는 공유 안 됨 © 숙대 창병모

61 dup2(3,1) 구현 1 2 34 Active inode table Open file table
Fd table (per process) Disk File access ... File size File type Inode list File pos refcnt=1 ... 1 2 34 File access ... File size File type File pos refcnt=1 ... Data blocks © 숙대 창병모

62 dup2(3,1) 구현 1 2 34 Active inode table Open file table
Fd table (per process) Disk File access ... File size File type Inode list File pos refcnt=1 ... 1 2 34 File access ... File size File type File pos refcnt=1 ... Data blocks © 숙대 창병모

63 dup(), dup2() 의 용도 표준 입출력의 redirection 표준 입출력의 파일 디스크립터
표준 입출력 대상을 파일로 바꿈 표준 입출력의 파일 디스크립터 #include <unistd.h> #define STDIN_FILENO /* 표준 입력 */ #define STDOUT_FILENO 1 /* 표준 출력 */ #define STDERR_FILENO 2 /* 표준 에러 */ © 숙대 창병모

64 예제: /* dup1.c */ #include <unistd.h> #include <fcntl.h>
#include <stdio.h> int main() { int fd; if((fd = creat("afile", 0600)) == -1) perror("afile"); printf("This is displayed on the screen.\n"); dup2(fd, STDOUT_FILENO); printf("This is written into the redirected file.\n"); return 0; } © 숙대 창병모

65 예제: /* dup2.c */ #include <unistd.h> #include <fcntl.h>
#include <stdio.h> int main() { int fd; if((fd = creat("afile", 0600)) == -1) perror("afile"); printf("This is displayed on the screen.\n"); close(STDOUT_FILENO); dup(fd); printf("This is written into the redirected file.\n"); return 0; } © 숙대 창병모

66 open() 두 번 같은 파일 두 번 open(“file”, O_RDONLY); file position은 공유되지 않는다 1
Open file table Active inode table Fd table (per process) Disk Inode list 1 File pos refcnt=1 ... 2 34 File access ... File size File type Data blocks File pos refcnt=1 ... © 숙대 창병모

67 예제 /* open2.c */ #include <fcntl.h> /* open2.c */
#include <unistd.h> int main() { int fd1, fd2, n; char buf[11]; fd1 = open("afile", O_RDONLY); fd2 = open("afile", O_RDONLY); if (fd1 < 0 || fd2 < 0) perror("afile"); n = read(fd1, buf, 10); buf[n] = ‘\0’; puts(buf); n = read(fd2, buf, 10); close(fd1); close(fd2); return 0; } © 숙대 창병모

68 두 프로세스에서 같은 파일 open() 1 2 34 1 2 34 Fd table(per process)
Open file table Active inode table Disk 1 Inode list 2 34 File pos refcnt=2 ... File access ... File size File type 1 2 34 File pos refcnt=2 ... Data blocks © 숙대 창병모

69 fork() 1 2 34 1 2 34 Fd table (per process) Open file table
Active inode table Disk 1 Inode list 2 34 File pos refcnt=2 ... File access ... File size File type Data blocks 1 2 34 © 숙대 창병모


Download ppt "제2장. 파일 입출력 (File I/O) © 숙대 창병모."

Similar presentations


Ads by Google