PA3-Spring2012 - Shell Design CS252 Systems Programming Computer Science Department Purdue University Shell Project To interact with the OS you use

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Shell Design CS252: Systems Programming Computer Science Department Purdue University Shell Project To interact with the OS you use a shell program or command interpreter Csh C Shell Tcsh Enhanced C Shell Sh - Shell Ksh Korn Shell Bash GNU shell Windows Desktop Mac OS Finder X Windows Managers There are also other graphical shells like Shell Interpreter The shell project is divided into several subsystems: Parser: reads a command line and creates a command table. One entry corresponds to a component in the pipeline. Example: Command: ls al | grep me > file1 Command Table ls grep In:dflt -al me Out:file1 Err:dflt Shell Interpreter Executor: Creates new process for each entry in the command table. It also creates pipes to communicate the output of one process to the input of the next one. Also it redirects the stdinput, stdoutput, and stderr. A | b | c | d > out < in All pipe entries share the same stderr Shell Interpreter Other Subsystems Environment Variables: Set, Print, Expand env vars Wildcards: Arguments of the form a*a are expanded to all the files that match them. Subshells: Arguments with ``(back[cks) are executed and the output is sent as input to the shell. Shell Project Part 1: Shell Parser. Read Command Line and print Command Table Part 2: Executer: Create Processes and communicate them with pipes. Also do in/out/err redirec[on. Part 3: Other Subsystems: Wildcard, Envars, Subshells Lex and Yacc A parser is divided into a lexical analyzer that separates the input into tokens and a parser that parses the tokens according to a grammar. The tokens are described in a file shell.l using regular expressions. The grammar is described in a file shell.y using syntax expressions. Shell.l is processed with a program called lex that generates a lexical analyzer. Shell.y is processed with a program called yacc that generates a parser program Shell Project Final Command Table ls grep In:dflt -al me Out:file1 Err:dflt aab aaa Lexer characters Shell.l Parser shell.y Wildcard and Envars Command Table Executor ls al a* | grep me > file1 <ls> <al> <a*> <PIPE> <grep> <me> <GREAT> ls grep In:dflt -al me Out:file1 a* Err:dflt Shell Grammar You need to implement the following grammar in shell.l and shell.y cmd [arg]* [ | cmd [arg]* ]* [ [> filename] [< filename] [ >& filename] [>> filename] [>>& filename] ]* [&] Currently, the grammar implemented is very simple Examples of commands accepted by the new grammar. ls al ls al > out ls al | sort >& out awk f x.awk | sort u < infile > oueile & Lexical Analyzer Lexical analyzer separates input into tokens. Currently shell.l supports a reduced number of tokens Step 1: You will need to add more tokens needed in the new grammar that are not currently in shell.l file ">>" { return GREATGREAT; } "|" { return PIPE;} "&" { return AMPERSAND} Etc. Shell Parser Step 2. Add the token names to shell.y %token NOTOKEN, GREAT, NEWLINE, WORD, GREATGREAT, PIPE, AMPERSAND etc Shell Parser Rules Step 3. You need to add more rules to shell.y pipe_list cmd [arg]* [ | cmd [arg]* ]* cmd_and_args arg_list io_modifier_list [ [> filename] [< filename] [ >& filename] [>> filename] [>>& filename] ]* io_modifier [&] command_line background_optional Shell Parser Rules goal: command_list; arg_list: arg_list WORD | /*empty*/ ; cmd_and_args: WORD arg_list ; Shell Parser Rules pipe_list: pipe_list PIPE cmd_and_args | cmd_and_args ; Shell Parser Rules io_modifier: GREATGREAT Word | GREAT Word | GREATGREATAMPERSAND Word | GREATAMPERSAND Word | LESS Word ; Shell Parser Rules io_modifier_list: io_modifier_list io_modifier | /*empty*/ ; background_optional: AMPERSAND | /*empty*/ ; Shell Parser Rules command_line: pipe_list io_modifier_list background_opt NEWLINE | NEWLINE /*accept empty cmd line*/ | error NEWLINE{yyerrok;} /*error recovery*/ command_list : command_list command_line ;/* command loop*/ Shell Parser Rules This grammar implements the command loop in the grammar itself. The error token is a special token used for error recovery. error will parse all tokens un[l a token that is known is found like <NEWLINE>. Yyerrok tells parser that the error was recovered. You need to add ac[ons {...}in the grammar to fill up the command table. Example: arg_list: arg_list WORD{currsimpleCmd->insertArg($2);} | /*empty*/ ; The Open File Table The process table also has a list with all the files that are opened Each open file descriptor entry contain a pointer to an open file object that contains all the informa[on about the open file. Both the Open File Table and the Open File Objects are stored in the kernel. The Open File Table The system calls like write/read refer to the open files with an integer value called file descriptor or fd that is an index into the table. The maximum number of files descriptor per process is 32 by default but but it can be changed with the command ulimit up to 1024. The Open File Table Open File Table 0 1 2 3 4 . . 31 Open File Object I-NODE Open Mode Offset Reference Count Open File Object An Open File Object contains the state of an open file. I-Node It uniquely iden[fies a file in the computer. An I-nodes is made of two parts: Major number Determines the devices Minor number It determines what file it refers to inside the device. Open Mode How the file was opened: Read Only Read Write Append Open File Object Offset The next read or write opera[on will start at this offset in the file. Each read/write opera[on increases the offset by the number of bytes read/wrioen. It is increased by the number of file descriptors that point to this Open File Object. When the reference count reaches 0 the Open File Object is removed. The reference count is ini[ally 1 and it is increased aper fork() or calls like dup and dup2. Reference Count Default Open Files When a process is created, there are three files opened by default: 0 Default Standard Input 1 Default Standard Output 2 Default Standard Error Write(1, "Hello", 5) Sends Hello to stdout Write(2, "Hello", 5) Sends Hello to stderr Stdin, stdout, and stderr are inherited from the parent process. The open() system call int open(filename, mode, [permissions]), It opens the file in filename using the permissions in mode. Mode: O_RDONLY, O_WRONLY, O_RDWR, O_CREAT, O_APPEND, O_TRUNC O_CREAT If the file does not exist, the file is created.Use the permissions argument for ini[al permissions. Bits: rwx(user) rwx(group) rwx (others) Example: 0555 Read and execute by user, group and others. (101B==5Octal) O_APPEND. Append at the end of the file. O_TRUNC. Truncate file to length 0. See "man open" The close() System call void close(int fd) Decrements the count of the open file object pointed by fd If the reference count of the open file object reaches 0, the open file object is reclaimed. The fork() system call int fork() It is the only way to create a new process in UNIX The OS creates a new child process that is a copy of the parent. ret = fork() returns: ret == 0 in the child process ret == pid > 0 in the parent process. ret < 0 error The memory in the child process is a copy of the parent process's memory. We will see later that this is op[mized by using VM copy- on-write. The fork() system call The Open File table of the parent is also copied in the child. The Open File Objects of the parent are shared with the child. Only the reference counters of the Open File Objects are increased. The fork() system call Open FileTable (parent)_ 0 1 2 3 Before: Open File Object Ref count=1 Ref count=1 Ref count=1 The fork() system call A)er: Open File Object Ref count=2 Open FileTable (child) 0 1 2 3 Open FileTable (parent) 0 1 2 3 Ref count=2 Ref count=2 The fork() system call Implica[on of parent and child sharing file objects: By sharing the same open file objects, parent and child or mul[ple children can communicate with each other. We will use this property to be able to make the commands in a pipe line communicate with each other. The execvp() system call int execvp(progname, argv) Loads a program in the current process. The old program is overwrioen. progname is the name of the executable to load. argv is the array with the arguments. Argv[0] is the progname itself. The entry aper the last argument should be a NULL so execvp() can determine where the argument list ends. If successful, execvp() will not return. The execvp() system call Example: Run "ls al" from a program. void main() { // Create a new process int ret = fork(); if (ret == 0) { // Child process. // Execute "ls al" const char *argv[3]; argv[0]="ls"; argv[1]="-al"; argv[2] = NULL; execvp(argv[0], argv); // There was an error perror("execvp"); exit(1); } else if (ret < 0) { // There was an error in fork perror("fork"); exit(2); } else { // This is the parent process // ret is the pid of the child // Wait until the child exits waitpid(ret, NULL); } // end if }// end main The execvp() system call For lab3 part2 start by crea[ng a new process for each command in the pipeline and making the parent wait for the last command. Command::execute() { int ret; for ( int i = 0; i < _numberOfSimpleCommands; i++ ) { ret = fork(); if (ret == 0) { //child execvp(sCom[i]->_args[0], sCom[i]->_args); perror("execvp"); exit(1); } else if (ret < 0) { perror("fork"); return; } // Parent shell continue } // for if (!background) { // wait for last process waitpid(ret, NULL); } }// execute The dup2() system call int dup2(fd1, fd2) Aper dup2(fd1, fd2), fd2 will refer to the same open file object that fd1 refers to. The open file object that fd2 refered to before is closed. The reference counter of the open file object that fd1 refers to is increased. dup2() will be useful to redirect stdin, stdout, and also stderr. The dup2() system call Example: redirec3ng stdout to file "myfile" previously created. Before: 0 1 2 3 Open File Object Shell Console Ref count=3 File "myout" Ref count=1 The dup2() system call A)er dup2(3,1); Open File Object Shell Console 0 1 2 3 Ref count=2 File "myout" Ref count=2 Now every prinH will go to file "myout". Example: Redirec[ng stdout A program that redirects stdout to a file myoutput.txt int main(int argc,char**argv) { // Create a new file int fd = open("myoutput.txt", O_CREAT|O_WRONLY|O_TRUNC, 0664); if (fd < 0) { perror("open"); exit(1); } // Redirect stdout to file dup2(fd,1); // Now printf that prints // to stdout, will write to // myoutput.txt printf("Hello world\n"); } The dup() system call fd2=dup(fd1) dup(fd1) will return a new file descriptor that will point to the same file object that fd1 is poin[ng to. The reference counter of the open file object that fd1 refers to is increased. This will be useful to "save" the stdin, stdout, stderr, so the shell process can restore it aper doing the redirec[on. The dup() system call Before: 0 1 2 3 Open File Object Shell Console Ref count=3 The dup() system call A)er fd2 = dup(1) Open File Object Shell Console 0 1 2 3 Ref count=4 fd2 == 3 The pipe system call int pipe(fdpipe[2]) fdpipe[2] is an array of int with two elements. Aper calling pipe, fdpipe will contain two file descriptors that point to two open file objects that are interconnected. What is wrioen into one of the file descriptors can be read from the other one and vice versa. The pipe system call Before: 0 1 2 3 Open File Objects Shell Console Ref count=3 The pipe system call A)er running: int fdpipe[2]; pipe(fdpipe); 0 1 fdpipe[0]==3 2 fdpipe[1]==4 3 What is wriPen in 4 fdpipe[0] can be read from fdpipe[1] and vice versa. Open File Objects Shell Console Ref count=3 pipe0 Ref count=1 Pipe1 Ref count=1 Example of pipes and redirec[on A program "lsgrep" that runs "ls al | grep arg1 > arg2". Example: "lsgrep aa myout" lists all files that contain "aa" and puts output in file myout. int main(int argc,char**argv) { if (argc < 3) { fprintf(stderr, "usage:" "lsgrep arg1 arg2\n"); exit(1); } // Strategy: parent does the // redirection before fork() //save stdin/stdout int tempin = dup(0); int tempout = dup(1); //create pipe int fdpipe[2]; pipe(fdpipe); //redirect stdout for "ls" dup2(fdpipe[0],1); close(fdpipe[0]); Example of pipes and redirec[on // fork for "ls" int ret= fork(); if(ret==0) { // close file descriptors // as soon as are not // needed close(fdpipe[1]); char * args[3]; args[0]="ls"; args[1]="-al"; args[2]=NULL; execvp(args[0], args); // error in execvp perror("execvp"); exit(1); } //redirection for "grep" //redirect stdin dup2(fdpipe[1], 0); close(fdpipe[1]); //create outfile int fd=open(argv[2], O_WRONLY|O_CREAT|O_TRUNC, 0600); if (fd < 0){ perror("open"); exit(1); } //redirect stdout dup2(fd,1); close(fd); Example of pipes and redirec[on // fork for "grep" ret= fork(); if(ret==0) { char * args[2]; args[0]="grep"; args[1]=argv[1]; args[2]=NULL; execvp(args[0], args); // error in execvp perror("execvp"); exit(1); } // Restore stdin/stdout dup2(tempin,0); dup2(tempout,1); // Parent waits for grep // process waitpid(ret,NULL); printf("All done!!\n"); } // main Execu[on Strategy for Your Shell Parent process does all the piping and redirec[on before forking the processes. The children will inherit the redirec[on. The parent needs to save input/output and restore it at the end. Stderr is the same for all processes a | b | c | d > ouWile < infile Execu[on Strategy for Your Shell execute(){ //save in/out int tmpin=dup(0); int tmpout=dup(1); //set the initial input int fdin; if (infile) { fdin = open(infile,......); } else { // Use default input fdin=dup(tmpin); } int ret; int fdout; for(i=0;i<numsimplecommands; i++) { //redirect input dup2(fdin, 0); close(fdin); //setup output if (i == numsimplecommands-1){ // Last simple command if(outfile){ fdout=open(outfile,......); } else { // Use default output fdout=dup(tmpout); } } Execu[on Strategy for Your Shell else { // Not last //simple command //create pipe int fdpipe[2]; pipe(fdpipe); fdout=fdpipe[0]; fdin=fdpipe[1]; }// if/else // Redirect output dup2(fdout,1); close(fdout); // Create child process ret=fork(); if(ret==0) { execvp(scmd[i].args[0], scmd[i].args); perror("execvp"); exit(1); } } // for Execu[on Strategy for Your Shell //restore in/out defaults dup2(tmpin,0); dup2(tmpout,1); close(tmpin); close(tmpout); if (!background) { // Wait for last command waitpid(ret, NULL); } } // execute Notes about Shell Strategy The key point is that fdin is set to be the input for the next command. fdin is a descriptor either of an input file if it is the first command or a fdpipe[1] if it is not the first command. This example only handles pipes and in/out redirec[on You have to redirect stderr for all processes if necessary You will need to handle the "append" case Implemen[ng Wildcards in Shell I suggest to implement first the simple case where you expand wildcards in the current directory. In shell.y, where arguments are inserted in the table do the expansion. Implemen[ng Wildcards in Shell Before argument: WORD Aper argument: WORD { expandWildcardsIfNecessary($1); } ; { Command::_currentSimpleCommand->insertArgument($1); } ; Implemen[ng Wildcards in Shell void expandWildcardsIfNecessary(char * arg) { // Return if arg does not contain `*' or `?' if (arg has neither `*' nor `?' (use strchr) ) { Command::_currentSimpleCommand->insertArgument(arg); return; } Implemen[ng Wildcards in Shell // 1. Convert wildcard to regular expression // Convert "*" -> ".*" // "?" -> "." // "." -> "\." and others you need // Also add ^ at the beginning and $ at the end to match // the beginning and the end of the word. // Allocate enough space for regular expression char * reg = (char*)malloc(2*strlen(arg)+10); char * a = arg; char * r = reg; *r = `^'; r++; // match beginning of line while (*a) { if (*a == `*') { *r=`.'; r++; *r=`*'; r++; } else if (*a == `?') { *r=`.' r++;} else if (*a == `.') { *r=`\\'; r++; *r=`.'; r++;} else { *a=*r; r++;} a++; } *r=`$'; r++; *r=0;// match end of line and add null char Implemen[ng Wildcards in Shell // 2. compile regular expression char * expbuf = compile( reg, 0, 0 ); if (expbuf==NULL) { perror("compile"); return; } // 3. List directory and add as arguments the entries // that match the regular expression DIR * dir = opendir("."); if (dir == NULL) { perror("opendir"); return; } Implemen[ng Wildcards in Shell struct dirent * ent; while ( (ent = readdir(dir))!= NULL) { // Check if name matches if (advance(ent->d_name, expbuf) ) { // Add argument Command::_currentSimpleCommand-> insertArgument(strdup(ent->d_name)); } closedir(dir); } Note: This code is not complete and contains errors. The purpose of this code is only to guide your implementation. Sor[ng Directory Entries Shells like /bin/sh sort the entries matched by a wildcard. For example "echo *" will list all the entries in the current directory sorted. You will have to modify the wildcard matching as follows: Sor[ng Directory Entries struct dirent * ent; int maxEntries = 20; int nEntries = 0; char ** array = (char**) malloc(maxEntries*sizeof(char*)); while ( (ent = readdir(dir))!= NULL) { // Check if name matches if (advance(ent->d_name, expbuf) ) { if (nEntries == maxEntries) { maxEntries *=2; array = realloc(array, maxEntries*sizeof(char*)); assert(array!=NULL); } array[nEntries]= strdup(ent->d_name); nEntries++; } } Sor[ng Directory Entries closedir(dir); sortArrayStrings(array, nEntries); // Add arguments for (int i = 0; i < nEntries; i++) { Command::_currentSimpleCommand-> insertArgument(array[i])); } free(array); Wildcards and Hidden Files In UNIX invisible files start with "." like .login, .bashrc etc. In these files that start with ".", the "." should not be matched with a wildcard. For example "echo *" will not display "." and "..". To do this, you will add a filename that starts with "." only if the wildcard has a "." at the beginning. Wildcards and Hidden Files if (advance (...) ) { if (ent->d_name[0] == `.') { if (arg[0] == `.') add filename to arguments; } } else { add ent->d_name to arguments } } Subdirectory Wildcards Wildcards also match directories inside a path: Eg. echo /p*/*a/b*/aa* To match subdirectories you need to match component by component Subdirectory Wildcards /usr/lib/aa /usr/lib/a* /usr/l*/a* /usr/local/a* /usr/lib/abb /usr/local/axz /usr/local/a12 /u/loc/aq /u/loc/ar /u/lll/apple /u/lll/atp /unix/l34/a* /unix/lll/a* /unix/l34/abcd /unix/l34/a45 /unix/lll/ab /unix/lll/ap3 /u*/l*/a* /u/l*/a* /u/loc/a* /u/lll/a* /unix/l*/a* Subdirectory Wildcards Strategy: Write a func[on expandWildcard(prefix, suffix) where prefix- The path that has been expanded already. It should not have wildcards. suffix The remaining part of the path that has not been expanded yet. It may have wildcards. /usr/l*/a* prefix suffix The prefix will be inserted as argument when the suffix is empty expandWildcard(prefix, suffix) is ini[ally invoked with an empty prefix and the wildcard in suffix. Subdirectory Wildcards #define MAXFILENAME 1024 void expandWildcard(char * prefix, char *suffix) { if (suffix[0]== 0) { // suffix is empty. Put prefix in argument. ...->insertArgument(strdup(prefix)); return; } // Obtain the next component in the suffix // Also advance suffix. char * s = strchr(suffix, `/'); char component[MAXFILENAME]; if (s!=NULL){ // Copy up to the first "/" strncpy(component,suffix, s-suffix); suffix = s + 1; } else { // Last part of path. Copy whole thing. strcpy(component, suffix); suffix = suffix + strlen(suffix); } Subdirectory Wildcards // Now we need to expand the component char newPrefix[MAXFILENAME]; if ( component does not have `*' or `?') { // component does not have wildcards sprintf(newPrefix,"%s/%s", prefix, component); expandWildcard(newPrefix, suffix); return; } // Component has wildcards // Convert component to regular expression char * expbuf = compile(...) char * dir; // If prefix is empty then list current directory if (prefix is empty) dir ="."; else dir=prefix; DIR * d=opendir(dir); if (d==NULL) return; Subdirectory Wildcards // Now we need to check what entries match while ((ent = readdir(d))!= NULL) { // Check if name matches if (advance(ent->d_name, expbuf) ) { // Entry matches. Add name of entry // that matches to the prefix and // call expandWildcard(..) recursively sprintf(newPrefix,"%s/%s", prefix, ent->d_name); expandWildcard(newPrefix,suffix); } } close(d); }// expandWildcard Execu[ng built-in func[ons All built-in func[ons except printenv are executed by the parent process. Why? we want setenv, cd etc to modify the state of the parent. If they are executed by the child, the changes will go away when the child exits. Execu[ng printenv Execute printenv in the child process and exit. In this way the output of printenv could be redirected. ret = fork(); if (ret == 0) { if (!strcmp(argument[0],printenv)) { char **p=environ; while (*p!=NULL) { printf("%s",*p); p++; } exit(0); } ... execvp(...); ... Signals and the Lex Scanner The scanner implemented by lex calls getc() to get the next char for stdin. getc() calls the system call read(). System calls that block will return if a signal is received. The errno will be EINTR. Ctrl-c generates the SIGINT signal. A child that exits will generate SIGCHLD. In any of these signals getc(), will return 1 (EOF) and it will exit. Signals and the Lex Scanner To prevent a system call to be interrupted, use sigac[on when se{ng the signal handler and also set the flag SA_RESTART. SA_RESTART will retry the system call if interrupted by a signal struct sigaction signalAction; signalAction.sa_handler = sigIntHandler; sigemptyset(&signalAction.sa_mask); signalAction.sa_flags = SA_RESTART; int error = sigaction(SIGINT, &signalAction, NULL ); if ( error ) { perror( "sigaction" ); exit( -1 ); } ...
View Full Document

This note was uploaded on 02/21/2012 for the course CS 252 taught by Professor Gustavorodriguez during the Spring '11 term at Purdue University-West Lafayette.

Ask a homework question - tutors are online