1. 1 IT 252 Computer Organization and Architecture Interaction Between Systems: File Sharing R. Helps

2. 2 System Interaction  As usual this can happen on many levels —Motherboard level: through buses —OS and app level: Files and file-sharing —Distributed system level: Networking Databases Cloud  Useful form of interaction is through shared file access  This discussion will focus on file access (chapter 10 of CS:APP)

3. 3 Sidenote: “Cloud” Basics  “Cloud” = remote shared computer services —Many existing technologies such as remote storage systems and remote databases are now labeled cloud services —Term “cloud” is generic and is being defined by marketplace  There are a few general types of cloud services —IaaS (Infrastructure as a Service) Remote physical or virtual machines Managed completely by the customer —PaaS (Platform as a Service) Remote system includes OS, database, web server (typically) Service provider maintains OS etc. (updates (etc) Customer develops on platform —SaaS (Software as a Service) Application packages provided as a service —Other variants available  General benefits of cloud: remote access, scale computing resources up and down very quickly and relatively cheaply (need extra servers for one week—rent from cloud)

4. 4 Files and IO  UNIX sees a file as a series of bytes  All IO in UNIX is seen as a file —Includes keyboards, displays, network connections, motors, sensors and anything else. —Some files can only be read (eg sensor, keypad) —Some files can only be written (eg display)  To use a file it must be opened, read, written, closed

5. 5 Open or Close files #include int fd; fd = open(char *filename, int flags, mode_t mode); close(fd); }  Returns fd, an integer, or (-1) if fail —fd = file descriptor Different from file pointer FILE *fp used with fopen —Points to a file in a process table  Filename is text filename  Flags are Read, Write, RW  Mode defines permissions (user, group, other) (use umask() );

6. 6 File Descriptors and read/write  File descriptors are allocated on open() and are reclaimed on close()  File descriptor point to file table  Once file is opened it can be read or written with read() & write() —Read()/write() keeps a pointer to current place in the file  If time discuss stat() and fstat() (file metadata) p 873 text

7. fd 0 fd 1 fd 2 fd 3 fd 4 Descriptor table (one table per process) Open file table (shared by all processes) v-node table (shared by all processes) File pos refcnt=1... File pos refcnt=1... stderr stdout stdin File access... File size File type File access... File size File type File A File B Data structure for open files

8. 8 File Descriptor assignment /* pp 10.1*/ #include "csapp.h" int main() { int fd1, fd2; fd1 = open(“foo.txt”, O_RDONLY, 0); close(fd1); fd2 = open(“baz.txt”, O_RDONLY, 0); printf(“fd2 = %d\n”, fd2); exit(0); }  What is the output?

9. 9 File Sharing  Multiple descriptors can point to same file  File are accessed through three data structures —Descriptor Table —File Table —V-node table  See 10.6 for details

10. fd 0 fd 1 fd 2 fd 3 fd 4 Descriptor table (one table per process) Open file table (shared by all processes) v-node table (shared by all processes) File pos refcnt=1... File pos refcnt=1... stderr stdout stdin File access... File size File type File A File B File Sharing

11. 11 File Sharing /* pp 10.2*/ #include "csapp.h" int main() { int fd1, fd2; char c; fd1 = open(“foobar.txt”, O_RDONLY, 0); fd2 = open(“foobar.txt”, O_RDONLY, 0); read(fd1, &c, 1); read(fd2, &c, 1); printf(“c = %c\n”, c); exit(0); }  Assume foobar.txt contains six characters “ foobar ”. What is the output?

12. fd 0 fd 1 fd 2 fd 3 fd 4 Descriptor table (one table per process) Open file table (shared by all processes) v-node table (shared by all processes) File pos refcnt=0... File pos refcnt=2... File access... File size File type File access... File size File type File A File B ??

13. Child-Parent fork  An app can fork a child process using fork()  The child process inherits the parent’s descriptor table —Has access to parent’s data 13

14. fd 0 fd 1 fd 2 fd 3 fd 4 Descriptor tables Open file table (shared by all processes) v-node table (shared by all processes) File pos refcnt=2... File pos refcnt=2... Parent's table fd 0 fd 1 fd 2 fd 3 fd 4 Child's table File access... File size File type File access... File size File type File A File B Child inherits Parent’s Open Files

15. 15 Parent-child sharing /* pp 10.3*/ #include "csapp.h" int main() { int fd; char c; fd = open(“foobar.txt”, O_RDONLY, 0); if (fork() == 0) { read(fd, &c, 1); exit(0) } wait(NULL); read(fd, &c, 1); printf(“c = %c\n”, c); exit(0); }  Assume foobar.txt contains six characters “ foobar ”. What is the output?

16. Sharing without child process  File descriptors can be duplicated  Share same open file description  Share file offset data  Use dup() or dup2() 16

17. 17 Dup() /* pp 10.5*/ #include "csapp.h" int main() { int fd1, fd2; char c; fd1 = open(“foobar.txt”, O_RDONLY, 0); fd2 = open(“foobar.txt”, O_RDONLY, 0); read(fd2, &c, 1); dup2(fd2, fd1); read(fd1, &c, 1); printf(“c = %c\n”, c); exit(0); }  Assume foobar.txt contains six characters “ foobar ”. What is the output?

18. Unix I/O functions (accessed via system calls) Standard I/O functions C application program fopen fdopen fread fwrite fscanf fprintf sscanf sprintf fgets fputs fflush fseek fclose open read write lseek stat close rio_readn rio_writen rio_readini tb rio_readlin eb rio_readnb RIO functions Memory-mapped I/O