CS 3214
Computer Systems
Godmar Back
PROCESSES
Part 1
CS 3214 Spring 2020
Processes
• Def: An instance of a program in execution
• OS provides each process with key abstractions– Logical control flow
• 1 flow – single-threaded process
• Multiple flows – multi-threaded process
– Private address space
– Abstracted resources: e.g., stdout/stdin file descriptors
• These abstractions create the illusion that each process has access to its own– CPU (or CPUs for multi-threaded processes)
– Memory
– Devices: e.g., terminal
CS 3214 Spring 2020
Context Switching
• Historical motivation for processes was introduction of multi-programming: – Load multiple processes into memory, and switch to
another process if current process is (momentarily) blocked
– This required protection and isolation between these processes, implemented by a privileged kernel: dual-mode operation.
• Time-sharing: switch to another process periodically to make sure all processes make equal progress
• Switch between processes is called a context switch
CS 3214 Spring 2020
CS 3214 Spring 2020
Dual-Mode Operation
• Two fundamental modes:– “kernel mode” – privileged
• aka system, supervisor or monitor mode
• Intel calls its PL0, Privilege Level 0 on x86
– “user mode” – non-privileged• PL3 on x86
• Bit in CPU – controls operation of CPU– Privileged operations can only
be performed in kernel mode. Example: hlt
– Must carefully control transition between user & kernel mode
int main()
{
asm(“hlt”);
}
Mode Switching
• User → Kernel mode– For reasons external or internal to CPU
• External (aka hardware) interrupt: – timer/clock chip, I/O device, network card, keyboard, mouse
– asynchronous (with respect to the executing program)
• Internal interrupt (aka software interrupt, trap, or exception)– are synchronous
– can be intended (“trap”): for system call (process wants to enter kernel to obtain services)
– or unintended (usually): (“fault/exception”) (division by zero, attempt to execute privileged instruction in user mode, memory access violation, invalid instruction, alignment error, etc.)
• Kernel → User mode switch on iret instruction
CS 3214 Spring 2020
CS 3214 Spring 2020
A Context Switch Scenario
Process 1
Process 2
Kernel
user mode
kernel mode
Timer interrupt: P1 is preempted,
context switch to P2
System call: (trap):
P2 starts I/O operation, blocks
context switch to process 1
I/O device interrupt:
P2’s I/O complete
switch back to P2
Timer interrupt: P2 still has
time left, no context switch
CS 3214 Spring 2020
Context Switching, Details
Process 1
Process 2
Kernel
user mode
kernel mode
intr_entry:
(saves entire CPU state)
(switches to kernel stack) intr_exit:
(restore entire CPU state)
(switch back to user stack)
iret
switch_threads: (in)
(saves caller’s state)
switch_threads: (out)
(restores caller’s state)(kernel stack switch)
CS 3214 Spring 2020
System Calls
Process 1
Kernel
user mode
kernel mode
User processes access kernel services by
trapping into the kernel, executing kernel
code to perform the service, then returning –
very much like a library call.
Unless the system call cannot complete
immediately, this does not involve a context
switch.
Kernel’s System Call Implementation
Syscall example: write(2)
• 32-bit Linux
CS 3214 Spring 2020
/* gcc -static -O -g -Wall write.c -o write */
#include <unistd.h>
int
main()
{
const char msg[] = "Hello, World\n";
return write(1, msg, sizeof msg);
}
0805005a <__write_nocancel>:
805005a: 53 push %ebx
805005b: 8b 54 24 10 mov 0x10(%esp),%edx #arg2
805005f: 8b 4c 24 0c mov 0xc(%esp),%ecx # arg1
8050063: 8b 5c 24 08 mov 0x8(%esp),%ebx # arg0
8050067: b8 04 00 00 00 mov $0x4,%eax # syscall no
805006c: cd 80 int $0x80
805006e: 5b pop %ebx
805006f: 3d 01 f0 ff ff cmp $0xfffff001,%eax
8050074: 0f 83 56 1e 00 00 jae 8051ed0 <__syscall_error>
805007a: c3 ret
/usr/include/asm/unistd.h:
….
#define __NR_write 4
….
CS 3214 Spring 2020
Kernel
Threads
Process 1
Process 2
Kernel
user mode
kernel mode
Most OS support kernel threads that never run in
user mode – these threads typically perform book
keeping or other supporting tasks. They do not
service system calls or faults.
Kernel Thread
Careful: “kernel thread” not the same as
kernel-level thread (KLT) – more on KLT later
Context vs Mode Switching
• Mode switch guarantees kernel gains control when needed– To react to external events
– To handle error situations
– Entry into kernel is controlled
• Not all mode switches lead to context switches– Kernel decides when – subject of scheduling policies
• Mode switch does not change the identity of current process/thread– See blue/yellow colors in slide on ctxt switch details
• Hardware knows about modes, does not (typically) know about contexts
CS 3214 Spring 2020
Bottom Up View: Exceptions
• An exception is a transfer of control to the
OS in response to some event (i.e., change
in processor state)
User Process OS
exception
exception processing
by exception handler
exception
return (optional)
event currentnext
CS 3214 Spring 2020
CS 3214 Spring 2020
Reasoning about Processes:
Process States
• Only 1 process (per CPU) can be in RUNNING state
RUNNING
READYBLOCKED
Process
must wait
for event
Event arrived
Scheduler
picks process
Process
preempted
Process States
• RUNNING:– Process is on CPU, its instructions are executed
• READY:– Process could make progress if a CPU were available
• BLOCKED:– Process cannot make progress even if a CPU were available
because it’s waiting for something (e.g., a resource, a signal, a point in time, a child to terminate, I/O, …)
• Model is simplified– OS have between 5 and 10 states typically
• Terminology not consistent across OS:– E.g., Linux calls BLOCKED “SLEEPING” and both READY and
RUNNING processes are called “RUNNING”; a “RUNNING” process is also called the ‘current’ process on its CPU.
CS 3214 Spring 2020
User View
• If process’s lifetimes overlap, they are said to
execute concurrently
– Else they are sequential
• Default assumption is concurrent execution
• Exact execution order is unpredictable
– Programmer should never make any assumptions
about it
• Any interaction between processes must be
carefully synchronized
CS 3214 Spring 2020
CS 3214 Spring 2020
Process Creation
• Two common paradigms:
– Cloning vs. spawning
• Cloning: (Unix)
– “fork()” clones current process
– child process then loads new program
• Spawning: (Windows)
– “exec()” spawns a new process with new program
• Difference is whether creation of new process
also involves a change in program
CS 3214 Spring 2020
fork()#include <unistd.h>
#include <stdio.h>
int
main()
{
int x = 1;
if (fork() == 0) {
// only child executes this
printf("Child, x = %d\n", ++x);
} else {
// only parent executes this
printf("Parent, x = %d\n", --x);
}
// parent and child execute this
printf("Exiting with x = %d\n", x);
return 0;
}
Child, x = 2
Exiting with x = 2
Parent, x = 0
Exiting with x = 0
CS 3214 Spring 2020
fork()#include <sys/types.h>
#include <unistd.h>
#include <stdio.h>
int main(int ac, char *av[])
{
pid_t child = fork();
if (child < 0)
perror(“fork”), exit(-1);
if (child != 0) {
printf ("I'm the parent %d, my child is %d\n",
getpid(), child);
wait(NULL); /* wait for child (“join”) */
} else {
printf ("I'm the child %d, my parent is %d\n",
getpid(), getppid());
execl("/bin/echo", "echo", "Hello, World", NULL);
}
}
fork/exec/exit/wait
CS 3214 Spring 2020
fork() wait()
exit()exec()
fork() vs. exec()
• fork():
– Clone most state of parent, including memory
– Inherit some state, e.g. file descriptors
– Keeps program, changes process
– Called once, returns twice
• exec():
– Overlays current process with new executable
– Keeps process, changes program
– Called once, does not return (if successful)
CS 3214 Spring 2020
exit(3) vs. _exit(2)
• exit(3) destroys current processes
• OS will free resources associated with it– E.g., closes file descriptors, etc. etc.
• Can have atexit() handlers– _exit(2) skips them
• Exit status is stored and can be retrieved by parent– Single integer
– Convention: exit(EXIT_SUCCESS) signals successful execution, where EXIT_SUCCESS is 0
CS 3214 Spring 2020
wait() vs waitpid()
• int wait(int *status)
– Blocks until any child exits
– If status != NULL, will contain value child passed to exit()
– Return value is the child pid
– Can also tell if child was abnormally terminated
• int waitpid(pid_t pid, int *status, int options)
– Can say which child to wait for
CS 3214 Spring 2020
If multiple children completed, will take in arbitrary order
– Can use macros WIFEXITED and WEXITSTATUS to get information about exit status
void fork10()
{
pid_t pid[N];
int i;
int child_status;
for (i = 0; i < N; i++)
if ((pid[i] = fork()) == 0)
exit(100+i); /* Child */
for (i = 0; i < N; i++) {
pid_t wpid = wait(&child_status);
if (WIFEXITED(child_status))
printf("Child %d terminated with exit status %d\n",
wpid, WEXITSTATUS(child_status));
else
printf("Child %d terminate abnormally\n", wpid);
}
}
Wait Example
CS 3214 Spring 2020
Observations on fork/exit/wait• Process can have many children at any point in time
• Establishes a parent/child relationship– Resulting in a process tree
• Zombies: processes that have exited, but their parent hasn’t waited for them– “Reaping a child process” – call wait() so that zombie’s
resources can be destroyed
• Orphans: processes that are still alive, but whose parent has already exited (without waiting for them)– Become the child of a dedicated process (“init”) who will
reap them when they exit
• “Run Away” processes: processes that (unintentionally) execute an infinite loop and thus don’t call exit() or wait()
CS 3214 Spring 2020
CS 3214 Spring 2020
The fork()/join() paradigm
• After fork(), parent & child execute in parallel– Unlike a fork in the road, here we
take both roads
• Used in many contexts
• In Unix, ‘join()’ is called wait()
• Purpose:– Launch activity that can be done in
parallel & wait for its completion
– Or simply: launch another program and wait for its completion (shell does that)
Parent:
fork()
Parent:
join()
Parent
process
executes
Child
process
executes
Child
process
exits
OS notifies
What do these
command lines do?
a) unix> one
b) unix> one first second third
c) unix> one &
d) unix> one < a
e) unix> one > b
f) unix> one | two
g) unix> one < a | two > b
h) unix> one | two | three | four &
i) unix> one & two & three
CS 3214 Spring 2020
FILE DESCRIPTORS
CS 3214 Spring 2020
Unix File Descriptors
• Unix provides a file descriptor abstraction
• File descriptors are– Small integers that have a local meaning within
one process
– Can be obtained from kernel • Several functions create them, e.g. open()
– Can refer to various kernel objects (not just files)
– Can be passed to a standard set of functions:• read, write, close, lseek, (and more)
– Can be inherited when a process forks a child
CS 3214 Spring 2020
Examples
• 0-2 are initially assigned
– 0 – stdin
– 1 – stdout
– 2 – stderr
– But this assignment is not fixed – process can
change it via syscalls
• int fd = open(“file”, O_RDONLY);
• int fd = creat(“file”, 0600);
CS 3214 Spring 2020
Implementing I/O Redirection
• dup and dup2() system call
• pipes: pipe(2)
CS 3214 Spring 2020
dup2
CS 3214 Spring 2020
#include <stdio.h>
#include <stdlib.h>
// redirect stdout to a file
int
main(int ac, char *av[])
{
int c;
int fd = creat(av[1], 0600);
if (fd == -1)
perror("creat"), exit(-1);
if (dup2(fd, 1) == -1)
perror("dup2"), exit(-1);
while ((c = fgetc(stdin)) != EOF)
fputc(c, stdout);
}
The Big Picture
CS 3214 Spring 2020
Process 1
0
1
2
user view kernel view
Terminal
Deviceopen(“x”)3
Open
File
File
Descriptor x
dup2(3,0)
Process 2
0
1
2
3
close (3)
Final steps (not included in
animation): Parent does
close(3), leaves the child’s
stdin to be the only reference
to the file descriptor.
Once the child is done and
closes its stdin (or exits!)
OS closes file and removes
entry from Open File table.
The Big Picture
CS 3214 Spring 2020
Process 1
0
1
2
user view kernel view
Terminal
Deviceopen(“x”)3
Open
File
File
Descriptor x
4
File
Descriptor
open(“x”) close(4)
Opening the same file within one process yields
separate read/write offsets for each descriptor
- Compare to dup/dup2/fork which create a
second reference to the same file descriptor
with a shared offset/file pointer.
Reference Counting
• Multiple file descriptors may refer to same open file– Within the same process:
• fd = open(“file”); fd2 = dup(fd);
– Across related processes:• fd = open(“file”); fork();
• But one process can also open a file multiple times:– fd = open(“file”); fd2 = open(“file”);
– In this case, fd and fd2 have different read/write offsets
• In both cases, closing fd does not affect fd2
• Reference Counting at 2 Levels:– Kernel keeps track of how many processes refer to a file descriptor –
fork() and dup()/dup2() may add refs
– And keeps track of how many file descriptors refer to an open file across all processes. Ditto for other kernel objects such as pipes.
• close(fd) removes reference in current process only!
CS 3214 Spring 2020
Practical Implications
• Number of simultaneously open file descriptors per process is limited– Soft limit of 1024 on current Linux, for instance;
can be increased to up to 64K (hard limit) on rlogin machines
• Must make sure fd’s are closed when done– Else ‘open()’ may fail
• Number space is reused– “double-close” error may inadvertently close a
new file descriptor assigned the same number
CS 3214 Spring 2020
IPC via “pipes”
• A bounded buffer providing a stream of bytes flowing through
• Properties– Writer() can put data in pipe as long as there is space
• If pipe() is full, writer blocks until reader reads()
– Reader() drains pipe()• If pipe() is empty, readers blocks until writer writes
• Classic abstraction– Decouples reader & writer
– Safe – no race conditions
– Automatically controls relative progress – if writer produces data faster than reader can read it, it blocks – and OS will likely make CPU time available to reader() to catch up. And vice versa.
CS 3214 Spring 2020
Fixed Capacity Buffer
write() read()
CS 3214 Spring 2020
int main()
{
int pipe_ends[2];
if (pipe(pipe_ends) == -1)
perror("pipe"), exit(-1);
int child = fork();
if (child == -1)
perror("fork"), exit(-1);
if (child == 0) {
char msg[] = { "Hi" };
close(pipe_ends[0]);
write(pipe_ends[1], msg, sizeof msg);
} else {
char bread, pipe_buf[128];
close(pipe_ends[1]);
printf("Child said "); fflush(stdout);
while ((bread = read(pipe_ends[0], pipe_buf, sizeof pipe_buf)) > 0)
write(1, pipe_buf, bread);
}
}
pipe
Note: there is no race condition in
this code. No matter what the
scheduling order is, the message sent
by the child will reach the parent.
esh – extensible shell
• Open-ended assignment
• Encourage collaborative learning
– Run each other’s plug-ins
• Does not mean collaboration on your
implementation
• Secondary goals:
– Exposure to yacc/lex and exposure to OO-
style programming in C
CS 3214 Spring 2020
Big Picture Issues• State maintenance
– How to maintain an accurate depiction of the state of external entities subject to change outside of the program’s control, when…
– external entities can change state asynchronously, and …
– tools for monitoring state changes (e.g., signals) are imperfect
• Concurrency control– How to ensure the correctness of data when …
– the control flow is subject to asynchronous interruption (e.g., by signal handling), and…
– there are complex control flows in shell
CS 3214 Spring 2020
Big Questions … and what you need to know
• In what state is every child process of the shell? How can the shell affect the state of its child processes?
– How to create processes executing specified code
– Signals and their effects on the receiver
– How to send/receive signals
– Terminal control and I/O
• What functions in the shell code must not be interrupted by asynchronous events?
– System calls for receiver to manage incoming signals
– How to handle incoming signals
CS 3214 Spring 2020
Using the list
implementation
• Key features: “list cell” – here call ‘list_elem’ is embedded in each object being kept in list– Means you need 1 list_elem per list you want to keep an object in
CS 3214 Spring 2020
struct esh_pipeline:
….
struct list commands
struct list_elem head
struct list_elem tail
struct list_elem *next;
struct list_elem *prev;
struct list_elem *next;
struct list_elem *prev;
struct esh_command:
….
struct list_elem elem;
….
struct list_elem *next;
struct list_elem *prev;
struct esh_command:
….
struct list_elem elem;
….
struct list_elem *next;
struct list_elem *prev;
list_entry(e, struct esh_command, elem)
Unix Startup: Step 1
init [1]
[0] Process 0: handcrafted kernel process
Child process 1 execs /sbin/init
1. Pushing reset button loads the PC with the address of a small
bootstrap program.
2. Bootstrap program loads the boot block (disk block 0).3. Boot block program loads kernel binary (e.g., /boot/vmlinux)
4. Boot block program passes control to kernel.
5. Kernel handcrafts the data structures for process 0.
Process 0 forks child process 1
CS 3214 Spring 2020
Unix Startup: Step 2
init [1]
[0]
gettyDaemonse.g. sshd, httpd
/etc/inittabinit forks and execs
daemons per /etc/inittab, and forks
and execs a getty program
for the console
CS 3214 Spring 2020
Unix Startup: Step 3
init [1]
[0]
The getty process
execs a login
programlogin
CS 3214 Spring 2020
Unix Startup: Step 4
init [1]
[0]
login reads login and passwd.
if OK, it execs a shell.if not OK, it execs another getty
tcsh
CS 3214 Spring 2020
Shell Programs• A shell is an application program that runs
programs on behalf of the user.– sh – Original Unix Bourne Shell
– csh – BSD Unix C Shell, tcsh – Enhanced C Shell
– bash –Bourne-Again Shell
Execution is a sequence of read/evaluate steps
int main()
{
char cmdline[MAXLINE];
while (1) {
/* read */
printf("> ");
fgets(cmdline, MAXLINE, stdin);
if (feof(stdin))
exit(0);
/* evaluate */
eval(cmdline);
}
}
CS 3214 Spring 2020
void eval(char *cmdline)
{
char *argv[MAXARGS]; /* argv for execve() */
int bg; /* should the job run in bg or fg? */
pid_t pid; /* process id */
bg = parseline(cmdline, argv);
if (!builtin_command(argv)) {
if ((pid = fork()) == 0) { /* child runs user job */
if (execve(argv[0], argv, environ) < 0) {
printf("%s: Command not found.\n", argv[0]);
exit(0);
}
}
if (!bg) { /* parent waits for fg job to terminate */
int status;
if (waitpid(pid, &status, 0) < 0)
unix_error("waitfg: waitpid error");
}
else /* otherwise, don’t wait for bg job */
printf("%d %s", pid, cmdline);
}
} Simple Shell eval Function
CS 3214 Spring 2020
Problem with Simple Shell Example
• Shell correctly waits for and reaps foreground jobs.
• But what about background jobs?
– Will become zombies when they terminate.
– Will never be reaped because shell (typically) will not terminate.
– Creates a memory leak and prevent pids from being reused
• Solution: Reaping background jobs requires a mechanism called a signal.
• Asynchronous – can arrive at any time. OS will interrupt process as soon as it does
CS 3214 Spring 2020
CS 3214 Spring 2020
Signals• A signal is a small message that notifies a process
that an event of some type has occurred in the system.– Kernel abstraction for exceptions and interrupts of
many (unrelated) kinds.– Sent from the kernel to a process.– Different signals are identified by small integer ID’s– Signal usually carry along some information about why
they were sent.
ID Name Default Action Corresponding Event
2 SIGINT Terminate Interrupt from keyboard (ctl-c)
9 SIGKILL Terminate Kill program (cannot override or ignore)
11 SIGSEGV Terminate & Dump Segmentation violation
14 SIGALRM Terminate Timer signal
17 SIGCHLD Ignore Child stopped or terminated
CS 3214 Spring 2020
Sending Signals
• Sending a signal– Kernel sends a signal to a destination process by updating some
state in the context of the destination process. A signal that is sent but has not been delivered is said to be pending.
• Examples include:– Kernel has detected a system event such as divide-by-zero
(SIGFPE), or illegal memory access (SIGSEGV)
– The kernel wishes to inform a process about the termination of a child (SIGCHLD)
– Another process (or the process itself) has invoked the killsystem call to explicitly request a signal (SIGUSR1)
– The terminal driver relays a request by the user to interrupt (SIGINT) or suspend (SIGTSTP) a process in the foreground process.
– And others
CS 3214 Spring 2020
Receiving a signal
• A destination process receives a signal when it is forced by the kernel to react in some way to the delivery of the signal.
• Three possible ways to react (subject to the process’s discretion):– Ignore the signal (do nothing)
– Terminate the process.
– Catch the signal by executing a user-level function called a signal handler.
• Akin to a hardware exception handler being called in response to an asynchronous interrupt.
• Not all options apply to all signals, and different signals come with different defaults
CS 3214 Spring 2020
Default Actions
• Each signal type has a predefined default action, which is one of:
– The process terminates
– The process terminates and dumps core.
– The process stops until restarted by a SIGCONT signal.
– The process ignores the signal.
• Type ‘man 7 signal’ to learn what the default action for a given signal is.
CS 3214 Spring 2020
Pending Signals
• A signal is pending if it has been sent but not yet received
– There can be at most one pending signal of any particular type.
– A pending signal is received at most once.
– i.e., Signals are not queued• If a process has a pending signal of type k, then subsequent
signals of type k that are sent to that process are discarded.
• Example: [sigchlddoesnotqueue.c]
• A process can block the receipt of certain signals.
– Blocked signals can be sent, but will not be received until the signal is unblocked.
CS 3214 Spring 2020
Kernel Internals
• How does the kernel implement this?
• Kernel maintains pending and blocked bit vectors in the context of each process.– pending – represents the set of pending signals
• Kernel sets bit k in pending whenever a signal of type k is sent.
• Kernel clears bit k in pending whenever a signal of type k is received/delivered.
– blocked – represents the set of blocked signals
• Can be set and cleared by the application using the sigprocmask system call.
CS 3214 Spring 2020
Kernel Internals, cont’d
• When does the kernel deliver signals?– Only at certain delivery points
• If target process executes in user mode, delivery can occur at any time.– This is done by forcing the process to enter kernel mode, if
necessary via processor-to-processor interrupt
• If target process executes inside the kernel, delivery occurs only at the next point where kernel code checks for pending signals (so as to not complicate the kernel’s control flow)
• If target process is in the BLOCKED state (Linux: SLEEPING/INTERRUPTABLE), it is made READY and will check for pending signals.
CS 3214 Spring 2020
Kernel Internals, cont’d
• Suppose kernel is in kernel mode, and ready to deliver a signal
• Kernel computes pnb = pending & ~blocked
– The set of pending nonblocked signals for process p
• If (pnb == 0) – Pass control to next instruction in the logical flow for p.
– (in other words, don’t do anything)
• Else– Choose least nonzero bit k in pnb and force process p to
receive signal k.
– The receipt of the signal triggers some action by p
– Repeat for all nonzero k in pnb.
– Pass control to next instruction in logical flow for p.
CS 3214 Spring 2020
Examples
• The following slides show some examples
to demonstrate the usefulness and power
of the signal facility
CS 3214 Spring 2020
Sending Signals with kill Program• kill program
sends arbitrary signal to a process or process group
• Examples– kill –9 24818
• Send SIGKILL to process 24818
– kill –9 –24817
• Send SIGKILL to every process in process group 24817.
linux> ./forks 16
linux> Child1: pid=24818 pgrp=24817
Child2: pid=24819 pgrp=24817
linux> ps
PID TTY TIME CMD
24788 pts/2 00:00:00 tcsh
24818 pts/2 00:00:02 forks
24819 pts/2 00:00:02 forks
24820 pts/2 00:00:00 ps
linux> kill -9 -24817
linux> ps
PID TTY TIME CMD
24788 pts/2 00:00:00 tcsh
24823 pts/2 00:00:00 ps
linux>
CS 3214 Spring 2020
Sending Signals with kill Function
CS 3214 Spring 2020
void fork12()
{
pid_t pid[N];
int i, child_status;
for (i = 0; i < N; i++)
if ((pid[i] = fork()) == 0)
while(1); /* Child infinite loop */
/* Parent terminates the child processes */
for (i = 0; i < N; i++) {
printf("Killing process %d\n", pid[i]);
kill(pid[i], SIGINT);
}
/* Parent reaps terminated children */
for (i = 0; i < N; i++) {
pid_t wpid = wait(&child_status);
if (WIFEXITED(child_status))
printf("Child %d terminated with exit status %d\n",
wpid, WEXITSTATUS(child_status));
else
printf("Child %d terminated abnormally\n", wpid);
}
}
A Program That Reacts to
Externally Generated Events (ctrl-c)#include <stdlib.h>
#include <stdio.h>
#include <signal.h>
void handler(int sig) {
printf("You think hitting ctrl-c will stop the bomb?\n");
sleep(2);
printf("Well...");
fflush(stdout);
sleep(1);
printf("OK\n");
exit(0);
}
main() {
signal(SIGINT, handler); /* installs ctrl-c handler */
while(1) {
}
}
CS 3214 Spring 2020
A Program That Reacts to
Internally Generated Events#include <stdio.h>
#include <signal.h>
int beeps = 0;
/* SIGALRM handler */
void handler(int sig) {
printf("BEEP\n");
fflush(stdout);
if (++beeps < 5)
alarm(1);
else {
printf("BOOM!\n");
exit(0);
}
}
main() {
signal(SIGALRM, handler);
alarm(1); /* send SIGALRM in
1 second */
while (1) {
/* handler returns here */
}
}
linux> a.out
BEEP
BEEP
BEEP
BEEP
BEEP
BOOM!
bass>
CS 3214 Spring 2020
CS 3214 Spring 2020
static void
catch_segfault(int signo, siginfo_t *info, void * _ctxt)
{
ucontext_t * ctxt = _ctxt;
printf(
"Catching Segfault at sig=%d fault addr is %p eip was at %x\n",
signo, info->si_addr, ctxt->uc_mcontext.gregs[REG_RIP]);
// skip offending instruction
ctxt->uc_mcontext.gregs[REG_RIP] += 7;
// put 42 into %rsi
ctxt->uc_mcontext.gregs[REG_RSI] = 42;
// upon return, OS will apply any changes to ctxt
// to process's saved state, the restore the state
// execution will continue at 400668 + 7 = 40066f
}
int
main()
{
install_signal_handler(SIGSEGV, catch_segfault);
printf("Dereferencing NULL -> %d ...\n", *(int *)NULL);
return 0;
}
Catching
Segfaults
Terminal Related Signals
• Signals are also used to allow a user to
control the execution of jobs from a
terminal.
– E.g., ^C and ^Z
• This job control must be set up by the
shell, and occurs in a manner in which
kernel, shell, and terminal driver cooperate
• Talk about process groups first
CS 3214 Spring 2020
Motivation for Process Groups
• The key motivation for process groups was the need for a mechanism by which a signal can be sent to a group of processes
• E.g., ^C sends SIGINT to all processes spawned by a process running in the foreground
– Common misconception is that those are killed with ^C because they’re children of the foreground process.
• Rather, the shell must judiciously create process groups (and assign processes to them) to achieve the desired routing of SIGINT/SIGTSTP signals
CS 3214 Spring 2020
Process Groups• Every process
belongs to exactly one process group
Fore-
ground
job
Back-
ground
job #1
Back-
ground
job #2
Shell
Child Child
pid=10
pgid=10
Foreground
process group 20
Background
process group 32Background
process group 40
pid=20
pgid=20pid=32
pgid=32pid=40
pgid=40
pid=21
pgid=20
pid=22
pgid=20
getpgrp() – Return process
group of current process
setpgid() – Change process group of a process
CS 3214 Spring 2020
Process Groups, cont.
• Every process can form a new process group by declaring themselves a leader– setpgid(0, 0)
• But, this is not the only way – process groups can also be formed by having a parent place a process in its own or an existing group– E.g., shell places all children belonging to the same pipeline into the
same group
• Process groups are populated simply by adding processes to them– Restriction: process being added to process group must be part of the
same “session” – a concept that groups multiple procgroup’s.
• See /proc/<pid>/stat to learn pgid of a process
• Hint: in esh, call setpgid() both in shell and child:– Will not hurt to place a process into a pgroup it’s already in
– Avoid race condition that would occur if shell or child assumed that the respective other will have already done it
CS 3214 Spring 2020
Sending Signals from the Keyboard• Typing Ctrl-C (Ctrl-Z) sends a SIGINT (SIGTSTP) to every job in the
foreground process group.
– SIGINT – default action is to terminate each process
– SIGTSTP – default action is to stop (suspend) each process
Fore-
ground
job
Back-
ground
job #1
Back-
ground
job #2
Shell
Child Child
pid=10
pgid=10
Foreground
process group 20
Background
process
group 32
Background
process
group 40
pid=20
pgid=20pid=32
pgid=32
pid=40
pgid=40
pid=21
pgid=20
pid=22
pgid=20
CS 3214 Spring 2020
Stopped Processes
• Modern Unixes allow users to stop processes, which prevents them from being scheduled (even if READY) until the user changes their mind and continues them
• Can be modeled using additional states in standard process state model
– READY/STOPPED
– BLOCKED/STOPPED
CS 3214 Spring 2020
Stopped Processes
CS 3214 Spring 2020
BLOCKED
STOPPED
READY
STOPPED
BLOCKED
READY
RUNNING
(Mainly) under user control
(except for terminal-related SIGTTIN etc.)
Dependent on
occurrence of
event the process
is interested in
Managing Terminal Access
• What if multiple processes wish to read from the terminal?– (default behavior: undefined)
• Kernel uses process groups to arbitrate– One foreground process group per terminal
• Kernel will automatically suspend (via SIGTTIN) any process in a non-foreground process group that attempts to read from terminal
– Try: “vim &”
• Ditto if a process wants EXCLUSIVE write access to a terminal– E.g., vim does this, a normal write(1, …) does not
• It’s up to shell to manage access to the terminal– Learn when your background children were stopped because they attempted to read
from a terminal. Inform the user.
– Use tcsetpgrp() to set the terminal’s foreground process group when user says to put a job into the foreground
– Save and restore terminal state when making process group owner of terminal
• Example of resource arbitration
CS 3214 Spring 2020
Installing Signal Handlers• The signal function modifies the default action
associated with the receipt of signal signum:– sighandler_t *signal(int signum, sighandler_t *handler)
• Different values for handler:
– SIG_IGN: ignore signals of type signum
– SIG_DFL: revert to the default action on receipt of signals of type signum.
– Otherwise, handler is the address of a signal handler• Called when process receives signal of type signum
• Referred to as “installing” the handler.
• Executing handler is called “catching” or “handling” the signal.
• When the handler executes its return statement, control passes back to instruction in the control flow of the process that was interrupted by receipt of the signal.
CS 3214 Spring 2020
POSIX Signal Handling
• Instead of signal, use ‘sigaction()’
• Signal handler has slightly different signature as in signal()
– Provides additional functionality
• Recommended
– signal(2) is now obsolete
– In project, recommend you use esh_signal_sethandler which is a convenient wrapper for sigaction()
CS 3214 Spring 2020
esh code/* Install signal handler for signal 'sig' */
void
esh_signal_sethandler(int sig, sa_sigaction_t handler)
{
sigset_t emptymask;
sigemptyset(&emptymask);
struct sigaction sa = {
.sa_sigaction = handler,
.sa_mask = emptymask,
/* do not block any additional signals (besides
* ‘sig’) when signal handler is entered. */
.sa_flags = SA_RESTART | SA_SIGINFO
/* restart system calls when possible */
/* use three argument handler */
};
if (sigaction(sig, &sa, NULL) != 0)
esh_sys_fatal_error("sigaction failed for signal %d", sig);
}
CS 3214 Spring 2020
Living With Nonqueuing Signals• Must check for all terminated jobs
– Typically loop with waitpid(,..WNOHANG)
void child_handler2(int sig)
{
int child_status;
pid_t pid;
while ((pid = waitpid(-1, &child_status, WNOHANG)) > 0) {
/* update data structures
that child ‘pid’ changed status */
/* do not call printf() */
}
}
void fork15()
{
. . .
signal(SIGCHLD, child_handler2);
. . .
}
CS 3214 Spring 2020
Signals & Concurrency
• Signal handlers for external events can occur *anytime*– Unless blocked – must think of signal handler as concurrent
flow of control
CS 3214 Spring 2020
user mode
kernel mode
handler
regular program
signal delivered
Signal handler returns
sigreturn()
Concurrent Accesses To Data Structures
• Consider shell maintaining a list of jobs
– Main program forks, adds jobs
– SIGCHLD handler may reap jobs, perhaps
remove jobs from joblist
CS 3214 Spring 2020
void
list_insert (struct list_elem *before, struct list_elem *elem)
{
elem->prev = before->prev;
elem->next = before;
before->prev->next = elem;
before->prev = elem;
}
If signal arrives inside the
instructions doing the list
manipulation, signal handler will
see inconsistent list – calls to
list_insert will lead to havoc
Signals & Concurrency
• Blocking a signal guarantees that signal handler execution will not occur even when signal is delivered– Will occur as soon as the signal is unblocked
CS 3214 Spring 2020
user mode
kernel mode
handler
block(SIGNAL)
signal sent
Signal handler returns
sigreturn()
unblock(SIGNAL)signal pending
Reentrancy
• A function is said to be reentrant if it can be safely called again even while a call is still in progress (i.e., has not returned)– Could be on a regular control flow path, e.g. recursion
– Or 2nd call could be from signal handler
– Or (discuss this later in more detail) from another thread
• Examples of functions that are not reentrant– inet_ntoa(), strtok() – uses private buffer
– printf() – takes a lock
• You cannot call non-reentrant functions from a signal handler for signal ‘s’– Unless you prevent the delivery of ‘s’ during calls in your
main program via { block(s); ….; unblock(s); }
CS 3214 Spring 2020
Async-Signal Safety• ‘async-signal safe functions’ - safe to call from a
signal handler– Provide the signal is allowed to occur (i.e., is not
blocked) while calls to these functions are in progress – else no issue arises
• See list in man 2 signal. Includes waitpid(), etc.
• The kicker: printf() is not safe to call in a signal handler– Frequent source of bugs (even in some textbook
sample code….!)
– Can use ‘snprintf() + write(1, …)’ if needed
• Please read Web Aside ECF:SAFETY on Async-Signal Safety
CS 3214 Spring 2020
Avoiding Race Conditions in esh
• Identify data structures shared between signal handler and main program
– E.g., everything the signal handler (or functions called from it) accesses
• Then protect accesses to those data structures by blocking the signal around the access
• Use ‘assert()’ to verify that you did this correctly– assert(esh_signal_is_blocked(SIGCHLD));
• Aside: the technique of delaying such interrupts is used inside OS kernels in a very similar way, e.g. when devices trigger interrupts
CS 3214 Spring 2020
Signals & System Calls
• What if a signal becomes pending while the recipient is BLOCKED in a system call such as wait() or read()?
• This system call could take a long time to complete, but signals are intended to be a timely mechanism
• OS will resume (make READY) the process, deliver the signal, and then (under typical instructions, i.e. SIG_RESTART) restart the system call.– Nice – program need not be aware of this scenario
– But you’ll occasionally see it in strace
CS 3214 Spring 2020
Signals – Summary
• Universal mechanism to notify a process of events– Internal events (memory access violation, process-
internal timers, …)
– External events• User-driven: ^C, ^Z
• Resulting from other processes: explicit kill(2), or SIGCHLD
• Resulting from kernel event: e.g., SIGTTOU, SIGTTIN
– Process groups are vessels for the delivery of signals to an entire group
• Signal handler can change program state before returning– Extremely powerful
CS 3214 Spring 2020
CS 3214 Spring 2020
Nonlocal Jumps: setjmp/longjmp
• Powerful (but dangerous) user-level mechanism for transferring control to an arbitrary location.– Controlled way to break the procedure call/return discipline
– Useful for error recovery and signal handling
• int setjmp(jmp_buf j)
– Must be called before longjmp
– Identifies a return site for a subsequent longjmp.
– Called once, returns one or more times
• Implementation:– Remember where you are by storing the current register
context, stack pointer, and PC value in jmp_buf.
– Return 0
CS 3214 Spring 2020
setjmp/longjmp Example#include <setjmp.h>
jmp_buf buf;
main() {
if (setjmp(buf) != 0) {
printf("back in main due to an error\n");
else
printf("first time through\n");
p1(); /* p1 calls p2, which calls p3 */
}
...
p3() {
<error checking code>
if (error)
longjmp(buf, 1)
}
CS 3214 Spring 2020
setjmp/longjmp (cont)
• void longjmp(jmp_buf j, int i)
– Meaning:• return from the setjmp remembered by jump buffer j again...
• …this time returning i instead of 0
– Called after setjmp
– Called once, but never returns
• longjmp Implementation:
– Restore register context from jump buffer j
– Set %eax (the return value) to i
– Jump to the location indicated by the PC stored in jump buf j.
CS 3214 Spring 2020
A Program That Restarts Itself When ctrl-c’d
while(1) {
sleep(1);
printf("processing...\n");
}
}
Ctrl-c
Ctrl-c
Ctrl-c
CS 3214 Spring 2020
bass> a.out
starting
processing...
processing...
restarting
processing...
processing...
processing...
restarting
processing...
restarting
processing...
processing...
#include <stdio.h>
#include <signal.h>
#include <setjmp.h>
sigjmp_buf buf;
void handler(int sig) {
siglongjmp(buf, 1);
}
main() {
signal(SIGINT, handler);
if (!sigsetjmp(buf, 1))
printf("starting\n");
else
printf("restarting\n");
Limitation of setjmp/longjmp
• Longjmp restores stack pointer– Thus activates a new stack frame
– Stack frame must still be valid
• Consequence:– Can only longjmp “up the stack” to functions that haven’t yet
returned when longjmp() is called
– repositioning the stack pointer automatically “destroys” intermediate stack frames
• But does not call cleanup functions provided in some languages (e.g. C++ destructors or Java ‘finally’ clauses)
– Longjmp’ing “down the stack” would “reactivate” already destroyed stack frames
• Does not necessarily crash, but leads to unpredictable results
• Think of setjmp/longjmp as a low-level mechanism to implement a variant of C++/Java style exceptions
CS 3214 Spring 2020
Summary
• Signals provide process-level exception handling– Can generate from user programs
– Can define effect by declaring signal handler
• Some caveats– Very high overhead
• >10,000 clock cycles
• Only use for exceptional conditions
– Don’t have queues (exception: “real-time signals”)• Just one bit for each pending signal type
• Nonlocal jumps provide exceptional control flow within process– Within constraints of stack discipline
CS 3214 Spring 2020