“There are two ways of constructing a software design. One way is to make it so simple that there are obviously no deficiencies. The other way is to make it so complicated that there are no obvious deficiencies.”
C.A.R.Hoare
Processes and Memory
Copyright By PowCoder代写 加微信 powcoder
• Operating System (OS) only purpose is make the software run on the hardware
• OS is an overhead cost. Everything that the OS does requires resources from the hardware: memory, computation
• OS is a necessary abstraction for program writers so they don’t need to know hardware details
Processes and Memory
• OS manages all the memory for processes (execution state of a program), devices and communication (interrupts).
• OS will computationally solve many problems (search tasks) that programmer doesn’t have to worry about.
• To do this, OS also needs to have additional memory for each program and for itself.
• Memory contains both data and instructions (binary code).
Processes and Memory
• System memory is divided into two spaces
– only processes with certain privileges can r/w/x
– OS functions and data live here e.g. I/O, processes, devices
– protect the hardware by only accessing through this layer
– all user created processes privilege depends on who created
– These programs are treated as rogue/untrusted that can run and die
– processes in this space are independent
• User space processes use system calls to access Kernel space *
Processes and Memory
• Usercreatinganewprocess
– OS creates a new image of memory that will be used by
the process by cloning an existing process.
– This has permissions associated with r/w/x of user/group
• Thememoryinsidetheprocessisassignedavirtual address range
– Pieces of virtual memory get mapped to physical memory when they are needed during execution
Processes and Memory
Pointer to Arguments and Environment
Environment
§ Memory of a process is divided into several parts
§ A process can potentially have more memory than system supports
§ Large virtual memory
§ Size of a new process’ virtual memory address space varies with OS
process virtual
start_code 0x0
Application
0x40000000
stack_start
0xc0000000
0xc0000000
0xffffffff
Shared C library *.so
(supervisor) mode 1GB kernel components
Kernel virtual memory (code, data, heap, stack
User stack (created at run time)
Memory-mapped region for shared libraries
Run-time heap (created by malloc)
Read/write segment (.data, .bss)
Read-only code segment (.init, .text, .rodata)
Memory invisible to user code
%esp (stack pointer)
Loaded from the executable file
There is no strict format for the layout of a process
Processes and Memory
• Virtual memory of a process is mapped to physical memory.
• Physical memory is mixed: cache, RAM, disk, tape,
network…
One process
§ Multiple processes share the same finite memory resources
§ The physical primary memory (DRAM/SRAM) is easily exhausted
§ Whatever cannot fit is stored in secondary memory
§ OS does this management of virtual memory translation to physical memory (with some hardware help)
instr + data
Physical DRAM
OS specific
instr + data
instr + data
Initiating Processes
• the standard C library includes functions
that invoke Unix system calls
• a set of these functions allows you to initiate and manage the running of other programs or processes
• the shell uses these functions to start the programs that correspond to the commands you type or put into a script
• when a program is started the main function is called
int main(int argc, char *argv[], char *envp[])
– argc is the number of arguments passed
– argv is an array of pointers to strings containing the arguments
– envp is an array of pointers to strings containing the environment variables
Starting a program
• when the shell starts a command such as:
echo testing
it calls the main function with the arguments: argc = 2
argv[0] = “echo”
argv[1] = “testing”
envp[0] = “VARNAME=value” envp[1] = …
top of stack
On the stack:
HOME=/somewhere
PATH=……..
Initiating processes
• the following functions will start another
process: execl, execle, execv, execve • eg:
int execl(const char *path, const char *arg,
const char *arg,… (char *)0)
signifies the end of the list of pointers to arguments
• exec in its various forms switches the program execution to another program
• your program is terminated and the other program’s main function is called
• if exec is successful it doesn’t return
• if it does return, and the return result is negative, then the program was not found
• if it returns zero or greater, then the exec function itself has failed!
if(execl(“/bin/sort”, “sort”, “myfile”, (char *)0) == -1) {
perror(argv[0]);
exit(1); }
/*program should never reach this point*/
#include
#include
#include
int main(int argc, char *argv[]) {
if( execl(“/usr/bin/sort”, “sort”, “words.txt”, (char *)0) == -1) { perror(argv[0]);
system function for printing error messages after a system call error
exec does not create a new process
Source code for ./abcd
1 void main() {
2 printf(“a \n”);
3 printf(“b \n”);
5 execl(“./hello”, 0); 4 printf(“c \n”);
5 printf(“d \n”);
execl() successful? Yes
Source code for ./hello
1 void main() {
2 printf(“hello world\n”); 3 printf(“foo\n”);
4 printf(“bar\n”);
5 printf(“baz\n”);
execl() unsuccessful lines 4-6 executed
Parallel execution • exec is like a GOTO
– it jumps to another program and doesn’t return
• it is possible to start another program but still continue to execute using the fork function
Fork function
#include
pid_t fork(void);
• creates a child process that is a copy of the memory image of the parent
Fork function
• both the parent and the child programs run in parallel
• the return value from the fork function is different for the parent and the child
• fork returns:
– 0 in the child process
– the process id of the child in the parent process – -1 in the parent process if the fork failed
Fork function
• by checking the return value of the fork function the running program can determine if it is the parent or the child
if (!(result = fork())) if (!(result = fork())) {{
/* child in control */ /* child in control */
/* parent in control */ /* parent in control */ if (result < 0) if (result < 0) {{
printf(“fork failed”);
exit(1); }} ......
printf(“fork failed”); exit(1);
• usually, the child process will then use one of the exec functions to start a new program
• the parent continues to execute
• the parent can ignore the child or wait for it to exit
• there are Unix system functions that allow parents to control the child
Wait function
• the parent can wait until the child exits and get the exit value
#include
pid_t wait(int *status)
pid_t waitpid(pid_t pid, int *status, int options) 23
waits for any child process to exit
waits for a specific process to exit
Wait function
• the wait function returns the process id of the child process
• the exit value of the child can be extracted from the status value
• other information in the status value indicates if the child failed or was terminated (rather than terminated normally)
• Processes are an abstraction for the OS
• the exec system call functions allow you to start another program running but the parent is terminated
• the fork system call function will make a copy of a process and both parent and child processes will continue to execute
• the wait system call function allows a parent process to wait for a child process to exit
• picture acknowledgement: http://www-sop.inria.fr/epidaure/research.php
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com