[SOLVED] Questions about a C program...

trist007 · 07-27-2011, 08:26 PM

I'm studying an OS programming book, in particular network sockets. I have two questions(2nd one is down at the bottom). There is something I don't quite understand. What if this piece of code, which sets sigpipe's default action to ignore and opens a socket:

Code:

if ((u_ignore_sigpipe() != 0) || ((sock = socket(AF_INET, SOCK_STREAM, 0)) == -1))
   return -1;

is replaced by this

Code:

if  (((sock = socket(AF_INET, SOCK_STREAM, 0)) == -1) || (u_ignore_sigpipe() != 0) )
   return -1;

The books says that this problem would arise: "if u_ignore_sigpipe() fails, u_connect returns with an open file descriptor in sock. Since the calling program does not have the value of sock, this file descriptor could not be closed."

What I don't understand is that socket gets run first and returns a file descriptor to sock, then u_ignore_sigpipe() gets evaluated. So if u_ignore_sigpipe() fails then sock doesn't get assigned? It would make sense if in the if statement, order of precedence goes from right to left so that if u_ignore_sigpipe() fails, then the if statement would just break and socket() would not even run. Is this the case?

Here's the complete program

Code:

#include <ctype.h>
#include <errno.h>
#include <netdb.h>
#include <stdio.h>
#include <string.h>
#include <unistd.h>
#include <arpa/inet.h>
#include <sys/select.h>
#include <sys/socket.h>
#include <sys/types.h>
#include "uiciname.h"
#include "uici.h"

int u_ignore_sigpipe(void);

int u_connect(u_port_t port, char *hostn) {
   int error;
   int retval;
   struct sockaddr_in server;
   int sock;
   fd_set sockset;

   if (name2addr(hostn,&(server.sin_addr.s_addr)) == -1) {
      errno = EINVAL;
      return -1;
   }
   server.sin_port = htons((short)port);
   server.sin_family = AF_INET;

   if ((u_ignore_sigpipe() == -1) ||
        ((sock = socket(AF_INET, SOCK_STREAM, 0)) == -1))
      return -1;

   if (((retval =
       connect(sock, (struct sockaddr *)&server, sizeof(server))) == -1) &&
       ((errno == EINTR) || (errno == EALREADY))) {          /* asynchronous */
       FD_ZERO(&sockset);
       FD_SET(sock, &sockset);
       while (((retval = select(sock+1, NULL, &sockset, NULL, NULL)) == -1)
           && (errno == EINTR)) {
          FD_ZERO(&sockset);
          FD_SET(sock, &sockset);
       }
   }
   if (retval == -1) {
        error = errno;
        while ((close(sock) == -1) && (errno == EINTR));
        errno = error;
        return -1;
   }
   return sock;
}

I'm also confused about what happens if the connect() function fails with either EINTR or EALREADY. EINTR is just when the connect() function gets interrupted by a signal. EALREADY is when the connection request is already in progress on a socket. So if either one is true,

1. What does FD_ZERO(&sockset) and FD_SET(sock, &socket) do?
Do they zero out the file descriptor and then sets it back to value of sock?
2. Then it runs select(), what does that do exactly? The book only says that it calls select to detect that the file descriptor is ready for writing. So basically the TCP 3 way handshake is complete, and select() makes sure that the socket is open for write?

dwhitney67 · 07-27-2011, 08:50 PM

For your first question, in the C programming language, when a conditional statement has something like (A || B), if A is true, then B is not evaluated. B will only be evaluated when A is false.

Thus for the following statement:

Code:

if  (((sock = socket(AF_INET, SOCK_STREAM, 0)) == -1) || (u_ignore_sigpipe() != 0) )

if the socket is created successfully, then the u_ignore_sigpipe() function will NOT be called. (EDIT: This statement is incorrect; u_ignore_sigpipe() will be called if the socket is created successfully).

For your other question concerning FD_ZERO(), FD_SET(), etc., these work on a file-descriptor set. The former zeroes out the set, and the latter adds the socket descriptor to the set. When calling select(), the highest socket descriptor plus one (1), should be passed as the first arg to the function.

select() can be used to tell you when it is appropriate to read from a socket, write to it, or handle an error. select() can be used to monitor multiple sockets at a time. Typically it is used for determining when it is ok to read, but I've seen it used for purposes of determining when it is ok to write. I've never seen it used for determining errors.

If you want to learn Socket Programming with an easier guide, I recommend Beej's documentation.

trist007 · 07-27-2011, 11:00 PM

Great thank you dwhitney67.

dwhitney67 · 07-28-2011, 03:34 AM

Quote:

Originally Posted by trist007

Great thank you dwhitney67.

You're welcome, however I provided you with false information... sorry... I really shouldn't post replies late at night.

With the following statement:

Code:

if  (((sock = socket(AF_INET, SOCK_STREAM, 0)) == -1) || (u_ignore_sigpipe() != 0) )

I did not pay close enough attention to the first conditional; if the socket is created successfully, then the return value from socket() will NOT be -1; thus that first conditional will be FALSE. This will force the second half of the conditional to be evaluated; in other words, u_ignore_sigpipe() will be called.

Sorry for the confusion; the comment I made earlier regarding (A || B) still apply, but my analysis of the statement above was off.

Once again, sorry for the confusion.