System-level I/O

(video #1)

Alan L. Cox
[email protected]

Some slides adapted from CMU 15.213 slides

Objectives

Appreciate the ingenuity of UNIX I/O model

Be able to choose the right I/O interface for

the task

Be able to use that I/O interface robustly and

efficiently

Be able to perform I/O redirection and file

sharing between processes

Cox System-level I/O 2

Unix Files

A Unix file is a sequence of m bytes:

 B0, B1, ... , Bk , ... , Bm-1

All I/O devices are represented as files:

 /dev/fd/0 (stdin)
 /dev/sd1a (disk)
 /dev/tty (terminal)
Even the kernel is represented as a file:
 /dev/kmem (kernel memory image)
 /proc (kernel data structures)

Cox System-level I/O 3

Unix File Types
Regular file
 Binary or text file
 Unix does not know the difference!
Directory file
 A file that contains the names and locations of other files
Character special and block special files
 Terminals (character special) and disks (block special)
FIFO (named pipe)
 A file type used for interprocess communication
Socket
 A file type used for network communication between
processes

Cox System-level I/O 4

Unix I/O

The mapping of files to devices allows kernel

to export simple interface called Unix I/O
Key Unix idea: All input and output is handled
in a consistent and uniform way
Basic Unix I/O operations (system calls):
 Opening and closing files
• open() and close()
 Changing the current file position
• lseek()
 Reading and writing a file
• read() and write()

Cox System-level I/O 5

Opening Files
Opening a file informs the kernel that you are getting ready to access
that file

int fd; /* file descriptor */

if ((fd = open(“/etc/hosts”, O_RDONLY)) < 0) {
perror(“open”);
exit(1);
}

Returns a small identifying integer file descriptor (-1 on error)

Each process created by a Unix shell begins with three open files:
 0: standard input
 1: standard output
 2: standard error
Must specify mode:
 O_RDONLY
 O_WRONLY
 O_RDWR
Writable files need additional information (O_CREAT, O_TRUNC,
O_APPEND) and must also specify file permissions

Cox System-level I/O 6

Closing Files
Closing a file informs the kernel that you are
finished accessing that file
int fd; /* file descriptor */
if (close(fd) < 0) {
perror(“close”);
exit(1);
}

Returns 0 on success, -1 on failure

Closing an already closed file is a recipe for
disaster in threaded programs (more on this
later)
 Moral: Always check return codes, even for
seemingly benign functions such as close()

Cox System-level I/O 7

Reading Files
Reading a file copies bytes from the current file
position to memory, and then updates file position
char buf[512];
int fd; /* file descriptor */
ssize_t nbytes; /* number of bytes read */
/* Open file fd ... */
/* Then read up to 512 bytes from file fd */
if ((nbytes = read(fd, buf, sizeof(buf))) < 0) {
perror(“read”);
exit(1);
}

Returns number of bytes read from file fd into buf

 nbytes < 0 indicates that an error occurred
 short counts (nbytes < sizeof(buf) ) are possible and
are not errors!

Cox System-level I/O 8

Writing Files
Writing a file copies bytes from memory to the current
file position, and then updates current file position
char buf[512];
int fd; /* file descriptor */
ssize_t nbytes; /* number of bytes read */
/* Open the file fd ... */
/* Then write up to 512 bytes from buf to file fd */
if ((nbytes = write(fd, buf, sizeof(buf)) < 0) {
perror(“write”);
exit(1);
}

Returns number of bytes written from buf to file fd

 nbytes < 0 indicates that an error occurred
 As with reads, short counts are possible and are not
errors!

Cox System-level I/O 9

Unix I/O Example
Copying standard input to standard output one
byte at a time

#include "csapp.h"

int
main(void)
{
char c;

while(Read(STDIN_FILENO, &c, 1) != 0)
Write(STDOUT_FILENO, &c, 1);
return (0);
}

Note the use of error handling wrappers for read

and write

Cox System-level I/O 10

Dealing with Short Counts

Short counts can occur in these situations:

 Encountering (end-of-file) EOF on reads
 Reading text lines from a terminal
 Reading and writing network sockets or Unix pipes
Short counts never occur in these situations:
 Reading from disk files (except for EOF)
 Writing to disk files
How should you deal with short counts in your
code?
 Use the RIO (Robust I/O) package from csapp.c

Cox System-level I/O 11

The RIO Package
RIO is a set of wrappers that provide efficient and
robust I/O in applications such as network programs
that are subject to short counts

RIO provides two different kinds of functions

 Unbuffered input and output of binary data
• rio_readn and rio_writen
 Buffered input of text lines and binary data
• rio_readlineb and rio_readnb
• The buffered RIO routines are thread-safe and can be
interleaved arbitrarily on the same descriptor

Available at:
/clear/www/htdocs/comp321/src/csapp.c
/clear/www/htdocs/comp321/include/csapp.h

Cox System-level I/O 12

 Unbuffered RIO Input and Output

Same interface as Unix read and write

Especially useful for transferring data on
network sockets
#include “csapp.h”

ssize_t rio_readn(int fd, void *usrbuf, size_t n);

ssize_t rio_writen(int fd, void *usrbuf, size_t n);

Return: number of bytes transferred if OK, 0 on EOF (rio_readn only), -1 on error

 rio_readn returns a short count only if it

encounters EOF
 rio_writen never returns a short count
 Calls to rio_readn and rio_writen can be
interleaved arbitrarily on the same descriptor

Cox System-level I/O 13

Implementation of rio_readn
/* rio_readn - robustly read n bytes (unbuffered) */
ssize_t
rio_readn(int fd, void *usrbuf, size_t n)
{
size_t nleft = n;
ssize_t nread;
char *bufp = usrbuf;

while (nleft > 0) {

if ((nread = read(fd, bufp, nleft)) < 0) {
if (errno == EINTR) /* interrupted by sig
handler return */
nread = 0; /* so call read() again */
else
return (-1); /* errno set by read() */
}
else if (nread == 0)
break; /* EOF */
nleft -= nread;
bufp += nread;
}
return (n - nleft); /* return >= 0 */
}
Cox System-level I/O 14
Implementation of rio_writen
/* rio_writen - robustly write n bytes (unbuffered) */
ssize_t
rio_writen(int fd, void *usrbuf, size_t n)
{
size_t nleft = n;
ssize_t nwritten;
char *bufp = usrbuf;

while (nleft > 0) {

if ((nwritten = write(fd, bufp, nleft)) <= 0) {
if (errno == EINTR) /* interrupted by sig
handler return */
nwritten = 0; /* so call write() again */
else
return (-1); /* errno set by write() */
}
nleft -= nwritten;
bufp += nwritten;
}
return (n);
}

Cox System-level I/O 15

Buffered RIO Input Functions
Efficiently read text lines and binary data from a file
partially cached in an internal memory buffer
#include “csapp.h”

void rio_readinitb(rio_t *rp, int fd);

ssize_t rio_readlineb(rio_t *rp, void *usrbuf, size_t maxlen);

ssize_t rio_readnb(rio_t *rp, void *usrbuf, size_t n);

Return: number of bytes read if OK, 0 on EOF, -1 on error

 rio_readlineb reads a text line of up to maxlen bytes from
file fd and stores the line in usrbuf
• Especially useful for reading text lines from network sockets
 rio_readnb reads up to n bytes from file fd
 Calls to rio_readlineb and rio_readnb can be interleaved
arbitrarily on the same descriptor
• Warning: Don’t interleave with calls to rio_readn

Cox System-level I/O 16

 Why is it more efficient?

rio_t structure
ssize_t
rio_readnb(...)
{
char rio_buf[RIO_BUFSIZE]
while (nleft > 0) { refill rio_buf
when empty
suck data
} out of rio_buf
}
Hard
Harddisk
disk

Cox System-level I/O 17

RIO Example

Copying the lines of a text file from standard

input to standard output

#include "csapp.h"

int
main(int argc, char *argv[])
{
int n;
rio_t rio;
char buf[MAXLINE];

Rio_readinitb(&rio, STDIN_FILENO);
while((n = Rio_readlineb(&rio, buf, MAXLINE)) != 0)
Rio_writen(STDOUT_FILENO, buf, n);
return (0);
}

Cox System-level I/O 18

 System-level I/O
(video #2)

Alan L. Cox
[email protected]

Some slides adapted from CMU 15.213 slides

File Metadata
Metadata is data about data, in this case file data
Maintained by kernel, accessed by users with the stat
and fstat functions
/* Metadata returned by the stat and fstat functions */
struct stat {
dev_t st_dev; /* device */
ino_t st_ino; /* inode */
mode_t st_mode; /* protection and file type */
nlink_t st_nlink; /* number of hard links */
uid_t st_uid; /* user ID of owner */
gid_t st_gid; /* group ID of owner */
dev_t st_rdev; /* device type (if inode device) */
off_t st_size; /* total size, in bytes */
unsigned long st_blksize; /* blocksize for filesystem I/O */
unsigned long st_blocks; /* number of blocks allocated */
time_t st_atime; /* time of last access */
time_t st_mtime; /* time of last modification */
time_t st_ctime; /* time of last change */
};
Cox System-level I/O 20
Example of Accessing File Metadata
/* statcheck.c - Querying and manipulating a file’s meta data */
int
main(int argc, char *argv[]) unix% ./statcheck statcheck.c
{ type: regular, read: yes
struct stat stat; unix% chmod 000 statcheck.c
char *type, *readok; unix% ./statcheck statcheck.c
type: regular, read: no
Stat(argv[1], &stat);
if (S_ISREG(stat.st_mode)) /* file type*/
type = "regular";
else if (S_ISDIR(stat.st_mode))
type = "directory";
else
type = "other";
if ((stat.st_mode & S_IRUSR)) /* OK to read?*/
readok = "yes";
else
readok = "no";
printf("type: %s, read: %s\n", type, readok);
return (0);
}

Cox System-level I/O 21

Accessing Directories

The only recommended operation on

directories is to read its entries
#include <sys/types.h>
#include <dirent.h>
…
{
DIR *directory;
struct dirent *de;
…
if ((directory = opendir(dir_name)) == NULL)
error("Failed to open directory");
…
while ((de = readdir(directory)) != NULL)
printf("Found file: %s\n", de->d_name);
…
closedir(directory);
}

Cox System-level I/O 22

 How the Unix Kernel Represents Open
Files
Two descriptors referencing two distinct open files
Descriptor 1 (stdout) points to terminal, and descriptor
4 points to open disk file
Descriptor table Open file table v-node table
[one table per process] [shared by all processes][shared by all processes]
File A (terminal)
stdin fd 0 File access
stdout fd 1
File pos File size Info in
stderr fd 2 stat
fd 3 refcnt=1 File type struct

...
fd 4
...

File B (disk)
File access
File size
File pos
refcnt=1
File type

...
...

Cox System-level I/O 23

File Sharing
Two distinct descriptors sharing the same disk file
through two distinct open file table entries
 E.g., calling open twice with the same filename

Descriptor table Open file table v-node table

(one table (shared by (shared by
per process) all processes) all processes)
File A
fd 0 File access
fd 1
File pos File size
fd 2
fd 3 refcnt=1 File type
fd 4

...
...

File B

File pos
refcnt=1
...

Cox System-level I/O 24

How Processes Share Files

A child process inherits its parent’s open files

 Here is the situation immediately after a fork:

Descriptor Open file table v-node table

tables (shared by (shared by
all processes) all processes)
Parent's table File A
fd 0 File access
fd 1 File size
File pos
fd 2
fd 3 refcnt=2 File type

...
fd 4 ...

Child's table File B

File access
fd 0
fd 1 File size
File pos
fd 2 File type
fd 3 refcnt=2

...
...

fd 4

Cox System-level I/O 25

I/O Redirection
Question: How does a shell implement I/O
redirection?
 unix% ls > foo.txt
Answer: By calling the dup2(oldfd, newfd)
function
 Copies (per-process) descriptor table entry oldfd to
entry newfd
Descriptor table Descriptor table
before dup2(4,1) after dup2(4,1)

fd 0 fd 0
fd 1 a fd 1 b
fd 2 fd 2
fd 3 fd 3
fd 4 b fd 4 b
Cox System-level I/O 26
I/O Redirection Example

Before calling dup2(4,1), stdout (descriptor 1)

points to a terminal and descriptor 4 points
to an open disk file
Descriptor table Open file table v-node table
(one table (shared by (shared by
per process) all processes) all processes)
File A
stdin fd 0 File access
stdout fd 1 File size
File pos
stderr fd 2
fd 3 refcnt=1 File type

...
fd 4 ...

File B
File access
File size
File pos
refcnt=1
File type

...
...

Cox System-level I/O 27

I/O Redirection Example (cont)

After calling dup2(4,1), stdout is now

redirected to the disk file pointed at by
descriptor 4
Descriptor table Open file table v-node table
(one table (shared by (shared by
per process) all processes) all processes)
File A
fd 0 File access
fd 1 File size
File pos
fd 2
fd 3 refcnt=0 File type

...
fd 4
...

File B
File access
File size
File pos
refcnt=2
File type

...
...

Cox System-level I/O 28

I/O Redirection
/* tsh.c – tiny shell */

void eval(char *cmdline)

{
<…snip…>
/* Create a child process */
pid = Fork();

if (pid == 0) { /* Child Process */

<…snip…>

input_fd = Open(infile, O_RDONLY);

output_fd = Open(outfile, O_WRONLY|O_CREAT, S_IRUSR|S_IRGRP|S_IROTH);
dup2(input_fd, 0);
dup2(output_fd, 1);

/* Now load and run the program in the new job */

if (execve(argv[0], argv, environ) < 0) {
printf("%s: Command not found\n", argv[0]);
exit(0);
}
}
<…snip…>
}

Cox System-level I/O 29

Accessing the Terminal

Only one process group can be in control of

the terminal
 Foreground process has access
 All background processes do not
 Background processes receive SIGTTIN/SIGTTOU
signal if they attempt to read/write terminal
Must explicitly place child in the foreground to
allow it access to the terminal
 int tcsetpgrp(int filedes, pid_t pgid_id)
 Must block or ignore SIGTTOU while setting
foreground group (several other intricacies in
getting this right)

Cox System-level I/O 30

 System-level I/O
(video #2B)

Alan L. Cox
[email protected]

Some slides adapted from CMU 15.213 slides

Concurrent Unix I/O
How to deal with multiple I/O operations concurrently?
 For example: wait for a keyboard input, a mouse click and input
from a network connection
Select system call
int select(int n, fd_set *readfds, fd_set *writefds,
fd_set *exceptfds, struct timeval *timeout);
Poll system call (same idea, different implementation)
int poll(struct pollfd *ufds, unsigned int nfds, int timeout);

struct pollfd { int fd; /* file descriptor */

short events; /* requested events */
short revents; /* returned events */
};

Other faster mechanisms are also available

 /dev/poll (Solaris)
 kqueue (FreeBSD)
 /dev/epoll (Linux)
 POSIX real-time signals + sigtimedwait()

Cox System-level I/O 32

Standard I/O Functions

The C standard library (libc.a) contains a

collection of higher-level standard I/O
functions
 Presented earlier in the lab
Recall the standard I/O functions:
 Opening and closing files (fopen and fclose)
 Reading and writing bytes (fread and fwrite)
 Reading and writing text lines (fgets and fputs)
 Formatted reading and writing (fscanf and
fprintf)

Cox System-level I/O 33

 Unix I/O vs. Standard I/O vs. RIO

Standard I/O and RIO are implemented using

low-level Unix I/O

fopen fdopen
fread fwrite
fscanf fprintf
sscanf C application program
sprintf fgets rio_readn
fputs fflush rio_writen
fseek Standard I/O RIO
rio_readinitb
fclose functions functions
rio_readlineb
open read rio_readnb
Unix I/O functions
write lseek
(accessed via system calls)
stat close
mmap

Which ones should you use in your programs?

Cox System-level I/O 34

Pros and Cons of Unix I/O
Pros
 Unix I/O is the most general and lowest overhead
form of I/O
• All other I/O packages are implemented using Unix
I/O functions
 Unix I/O provides functions for accessing file
metadata
 No intermediate buffering (eliminates a copy)
Cons
 Dealing with short counts is tricky and error prone
 Efficient reading of text lines requires some form
of buffering, also tricky and error prone
 Both of these issues are addressed by the standard
I/O and RIO packages

Cox System-level I/O 35

Pros and Cons of Standard I/O

Pros:
 Buffering increases efficiency of small operations,
e.g., fgetc(), by decreasing the number of read and
write system calls
 Short counts are handled automatically
Cons:
 Provides no functions for accessing file metadata
 Standard I/O is not appropriate for input and
output on network sockets
 There are poorly documented restrictions on stdio
streams that interact badly with restrictions on
network sockets

Cox System-level I/O 36

Standard I/O Restrictions

Restrictions on streams:
 Restriction 1: input function cannot follow output
function without an intervening call to fflush,
fseek, fsetpos, or rewind
• Latter three functions all use lseek to change file
position
 Restriction 2: output function cannot follow an
input function without an intervening call to fseek,
fsetpos, or rewind
Restriction on sockets:
 You are not allowed to change the file position of a
socket

Cox System-level I/O 37

Standard I/O Workarounds
Workaround for restriction 1:
 Flush stream after every output
Workaround for restriction 2:
 Open two streams on the same descriptor, one for
reading and one for writing:
FILE *fpin, *fpout;

fpin = fdopen(sockfd, “r”);

fpout = fdopen(sockfd, “w”);

 However, this requires you to close the same descriptor

twice:

fclose(fpin);
fclose(fpout);

 Creates a deadly race in concurrent threaded programs!

Cox System-level I/O 38

Choosing I/O Functions
General rule: Use the highest-level I/O functions you
can
 Many C programmers are able to do all of their work
using the standard I/O functions
When to use standard I/O?
 When working with disk or terminal files
When to use raw Unix I/O
 When you need to fetch file metadata
 In rare cases when you need absolute highest
performance
When to use RIO?
 When you are reading and writing network sockets or
pipes, but RIO spins in a loop until operation completes,
thread does nothing else in the meantime
 Never use standard I/O or raw Unix I/O on sockets or
pipes, there are specialized functions for sockets to be
discussed

Cox System-level I/O 39

Next Time

Networking

Cox System-level I/O 40

A Typical Hardware System
CPU chip
register file

ALU
system bus memory bus

I/O main
bus interface
bridge memory

I/O bus Expansion slots for

other devices such
USB graphics disk as network adapters.
controller adapter controller

mousekeyboard monitor
disk
Cox System-level I/O 41
Reading a Disk Sector: Step 1
CPU chip
CPU initiates a disk read by writing
register file
a command, logical block number,
and destination memory address to
ALU
a port (address) associated with the
disk controller

main
bus interface
memory

I/O bus

USB graphics disk

controller adapter controller

mousekeyboard monitor
disk
Cox System-level I/O 42
Reading a Disk Sector: Step 2
CPU chip
register file Disk controller reads the sector and
performs a direct memory access
ALU (DMA) transfer into main memory

main
bus interface
memory

I/O bus

USB graphics disk

controller adapter controller

mousekeyboard monitor
disk
Cox System-level I/O 43
Reading a Disk Sector: Step 3
CPU chip
When the DMA transfer completes,
register file the disk controller notifies the CPU
with an interrupt (i.e., asserts a
ALU
special “interrupt” pin on the CPU)

main
bus interface
memory

I/O bus

USB graphics disk

controller adapter controller

mousekeyboard monitor
disk
Cox System-level I/O 44
Asynchronous I/O
POSIX P1003.4 Asynchronous I/O interface functions:
(available in Solaris, AIX, Tru64 Unix, Linux 2.6,…)
 aio_cancel
• cancel asynchronous read and/or write requests
 aio_error
• retrieve Asynchronous I/O error status
 aio_fsync
• asynchronously force I/O completion, and sets errno to ENOSYS
 aio_read
• begin asynchronous read
 aio_return
• retrieve return status of Asynchronous I/O operation
 aio_suspend
• suspend until Asynchronous I/O Completes
 aio_write
• begin asynchronous write
 lio_listio
• issue list of I/O requests

Cox System-level I/O 45