Proper way to break out of a recv() loop in C - c

I wrote a server/client program where I can copy files from client to server. The problem is that I can't break out of the while loop in my server when writing to the file is done. I can only open the new copied file when I close my server program because the file doesn't get closed with fclose(copyFile). The file gets copied successfully everytime. However it does work properly when I run the server/client on the same machine but when I move the client to another pc, the server keeps blocking on recv() in my server.
Server:
while (1)
{
int res = recv(s, buf, BUFSIZ, 0);
if (res == SOCKET_ERROR)
{
printf("3 error %d\n", WSAGetLastError);
break;
}
fwrite(buf, 1, res, copyFile);
if (strncmp(buf, "exit", 4) == 0)
{
break;
}
}
fclose(copyFile);
Client:
while (size == BUFSIZ)
{
size = fread(buf, 1, BUFSIZ, originalFile);
int r = send(ClientSocket, buf, size, 0);
if (r == SOCKET_ERROR)
{
printf("1 error: %d\n", WSAGetLastError);
}
}
int r = send(ClientSocket, "exit", 4, 0);
if (r == SOCKET_ERROR)
{
printf("2 error: %d\n", WSAGetLastError);
}
fclose(originalFile);
How can I properly exit the while() loop in my server?

You are attempting to indicate the end of the file with the word "exit", but there are at least two problems with that:
if the file being transferred contains the word "exit" then that could be misinterpreted as an end-of-file marker, and
you seem to be assuming that your send() calls at the client will be paired one-to-one with recv() calls at the server, so that the "exit" can be relied upon to appear at the beginning of the buffer when the server receives it. That is not a safe assumption.
(Note also that even if the server did happen to receive the "exit" at the beginning of the buffer, you would still write it to the file before recognizing it as an end-of-file marker. If this were the only issue then it would be easy to fix.)
You need a workable application-level protocol to help client and server communicate -- something that layers the needed additional structure on top of the flat byte stream that the socket connection provides. The "exit" terminator is an attempt at that, but it is not a workable solution, at least not by itself.
In contrast, consider HTTP: messages consist of a sequence of headers that are individually and as a group recognizable by their lexical form, followed by the message body. Among the things that the headers can convey is the length of the message body, and this is how the recipient can recognize the end of the message body without the sender closing the connection.
You do not need to implement HTTP, but you can be inspired by it. If you want to transmit messages with arbitrary length and content, without signaling the end of the message by closing the connection, then your best bet is to tell the recipient in advance how many bytes to expect in the message. That could be as simple as prepending a fixed-length message-length field.

Related

How to recv until theres nothing more to recv without eof?

So i need to recv an html file from the server to the client, the file is bigger than the buffer so i make several sends. Thats why i have this loop when i recv
while (i = recv(s, buf, TAM_BUFFER, 0)) {
if (i == -1) {
perror(argv[0]);
fprintf(stderr, "%s: error reading result\n", argv[0]);
exit(1);
}
while (i < TAM_BUFFER) {
j = recv(s, &buf[i], TAM_BUFFER - i, 0);
if (j == -1) {
perror(argv[0]);
fprintf(stderr, "%s: error reading result\n", argv[0]);
exit(1);
}
i += j;
}
/* Print out the file line by line. */
printf("%s", buf);
}
the send looks something like this:
while (fgets(buf, sizeof(buf), fp)){
if (send(s, buf, TAM_BUFFER, 0) != TAM_BUFFER) errout(hostname);
}
The problem is the loop never ends, becase it doesnt recv the eof and i is never 0, its just remain blocked there.
I cant do the close to send the eof because after he recv the whole file, the client will ask for another file.
I tryed to send a SIGALRM if the loop stays blocked for longer than 5 seconds but it doesnt work as expected, because the loop wont stop, and it will throw an error.
Also how can i do to be able to recv less than TAM_BUFFER?(in the send, change the TAM_BUFFER -> strlen(buf)) I know i need to change the interior loop, but then ill have the same problem, j will not be 0 never, so i dont know how could i end it.(or maybe i dont need the second loop in this case).
EDIT: i cant send the lenght of the file beucause of the protocol im following
TCP is a protocol used to transport a single unstructured octet stream in each direction. Shutdown of the connection (i.e. EOF) is the only way in TCP to signal to the peer that no more data will be sent in this connection. If you need a different way because you need to distinguish between multiple messages inside the same TCP connection then you need to use an application level protocol which can specify such message boundaries. This is usually done by fixed message size, prefixing the message with a length or by special boundary markers.
If you can't embed payload size in your protocol, you have to identify EOF by closing socket or checking for timeout. You can use select function and set timeout for it, see here Using select and recv to obtain a file from a web server through a socket and https://stackoverflow.com/a/30395738/4490542

How to send and receive newline character over sockets in a client server model?

I am trying to learn client server model in Linux and I have setup two C files namely server.c and client.c. These are the code snippets that I seem to have problems with.
server.c code snippet
char* message = "<query>\n";
write(client_socket_filedescriptor, message, sizeof(message));
client.c code snippet
char* message = "<query>\n";
read(socket_filedescriptor, buffer, sizeof(buffer));
printf("%s", buffer);
printf("\n\n");
printf("%s", message);
Now when I run my server and then when I run my client, I expect the printf statements to print the same strings that is <query>\n, but I keep getting different outputs for buffer and message variables.
The output looks a bit like this when I run client code.
Output image
As you see, these two strings are different. I am trying to simulate a typical TCP handshake and I want to make sure that these two strings are same and then client will start writing or doing something with that server. But I am having this trivial problem. Could anyone tell my how to resolve it? I plan to use strcmp to compare buffer and message variables, but as it stands now, strcmp doesn't return 0 since these are different strings afterall.
You are ignoring the count returned by read(). It can be -1, indicating an error, or zero, indicating end of stream, or a positive number, indicating how many bytes were received. You cannot assume that read() fills the buffer, or that a single send() or write() corresponds to a single recv() or read().
In detail:
write(client_socket_filedescriptor, message, sizeof(message));
You are only sending four bytes, the size of the pointer. And you're ignoring the return value. It should be
int count = write(client_socket_filedescriptor, message, strlen(message));
if (count == -1)
perror("write"); // or better
char* message = "<query>\n";
read(socket_filedescriptor, buffer, sizeof(buffer));
That should be
int count = read(socket_filedescriptor, buffer, sizeof(buffer));
if (count == -1)
perror("read"); // or better
else if (count == 0)
; // end of stream: the peer has disconnected: close the socket and stop reading
else
Back to your code:
printf("%s", buffer);
That should be
printf("%.*s", count, buffer);
I plan to use strcmp()
You should plan to use strncmp(), with count above as the length parameter. In any case you can't assume the input ends with a null unless you (a) ensure you send the null, which you aren't, and (b) write a read loop that stops when you've read it.

Knowing when Client has finished reading from the socket in C

I am writing a client-server program in c where I have to send multiple image files from the server to the client. Is there any way for the server to know when the client has finished reading from socket, creating the image file locally, and successfully written to it? P.S. I already tried sending a message to the socket and when I try to read the socket from the server, the program hangs. Any help will be much appreciated.
Here is a function from the Server code which sends the file to socket:
while(1)
{
unsigned char buf[256] = {0};
int n = fread(buf,1,256,fp);
if(n>0) { send(sockfd,buf,n,0); }
if(n<256) {
if(feof(fp))
printf("Sent to socket\n");
break;
}
}
fclose(fp);
}
char buf[5]
read(sockfd,buf,5);
if(strcmp(buf,"ready")==0) //send more files
And here is a function from the client to write to the file:
FILE* fp;
fp = fopen(file_path,"ab");
char buf[256];
int num;
int total=0;
while(( num = recv(sockfd,buf,256,0))>0)
{
total+=num;
fwrite(buf,1,num,fp);
}
fclose(fp);
write(sockfd,"ready",5);
}
When I do a read on the server after one file transfer, the program hangs.
You problem is here in the client:
while(( num = recv(sockfd,buf,256,0))>0)
recv() will only return 0 at end-of-file - ie when the server shuts down the sending side of the socket. However your server isn't doing that - it's waiting for a response from the client. This means you deadlock.
If you want to send multiple images in the one connection, you need to send your client some information to allow it to tell when one image ends. One way to do this is to first send your client the size of the file it should expect, then the file data. The client can then keep track of how many bytes it's recieved, and send the "ready" reponse after that.
You need a kind of protocol to allow :
the server to tell the client that all has been written
the client to tell the server that all has been read
If you only send one file, the simplest way is to use shutdown(sockfd, SHUT_WR) server side after all data has beed sent. That way the client will get a 0 as return from recv signaling end of data and will be able to send its acknowledgement. But you can no longer write on the socket server side.
If you want to be able to send more than one file, you will have to imagine a cleverer protocol. A common one would be to send blocs of data preceded by their size
short sz = htons(n); /* deals with possible endianness problems */
send(sockfd, &sz, sizeof(short), 0)
send(sockfd, buf, n);
And a sz == 0 (with no buf ...) would signal end of data
This would still be a simple protocol with no error recovery, but at least it can work when no incident happens.

C sockets, Incomplete file transfer

I'm writing a C program to transfer a file of fixed size, a little over 2Mb, from a server to a client. I'm using TCP sockets on Linux and the code I wrote is the following:
Server (sender)
while (1) {
int nread = read(file, buffer, bufsize);
if (nread == 0) // EOF
break;
if (nread < 0) {
// handle errors
}
char* partial = buffer;
while (nread > 0) {
int nwrite = write(socket, partial, nread);
if (nwrite <= 0) {
// handle errors
}
nread -= nwrite;
partial += nwrite;
}
}
// file sent
shutdown(socket, SHUT_WR);
Client (receiver)
while (filesize > 0) {
nread = read(socket, buffer, bufsize);
if (nread == 0) {
// EOF - if we reach this point filesize is still > 0
// so the transfer was incomplete
break;
}
if (nread < 0) {
// handle errors
}
char* partial = buffer;
while (nread > 0) {
nwrite = write(file, partial, nread);
if (nwrite <= 0) {
// handle errors
}
nread -= nwrite;
partial += nwrite;
filesize -= nwrite;
}
}
if (filesize > 0) {
// incomplete transfer
// handle error
}
close(socket);
When testing the code on my laptop (both client and server "are" on localhost and the communication happen on the loopback interface), sometimes the client exits because read received an EOF, and not because it received all filesize bytes. Since i use a shutdown on the server, this should mean that there is no other data to read.
(Note that the server sent all the bytes and executed the shutdown correctly)
Can you explain me why this is happening?
Where are the missing bytes gone?
-----
EDIT 1 - Clarifications
Some users asked a couple of clarifications so i am posting the answers here:
The program is using TCP blocking sockets
The filesize is a fixed value and is hardcoded in both client and server.
No special socket options as, for example, SO_LINGER are enabled/used.
When the error occur, the server (sender) has already sent all the data and executed the shutdown correctly.
The error, as of today, never happened when testing the application with the client and the server on different machines (transfer over a real network interface and not a loopback interface)
EDIT 2
User Cornstalks pointed me to a really interesting article about the, non always reliable, behaviours of TCP.
The article, which is well worth a read, describe a few tricks useful when sending an unknown amount of data between TCP sockets. The tricks described are the followings:
Take advantage of the SO_LINGER option on the sender. This will help to keep the socket open, upon a call to close(2) or shutdown(2), until all the data has successfully been sent.
On the receiver, beware of pending readable data before the actual receiving loop. It could lead to an immediate reset being sent.
Take advantage of shutdown(2) to signal the receiver the the sender has done sending data.
Let the receiver know the size of the file that will be sent before actually sending the file.
Let the receiver acknowledge the sender that the receiving loop is over. This will help to prevent the sender from closing the socket too soon.
After reading the article, i upgraded my code to implement the tricks number 1 and 5.
This is how i implemented trick number 5:
Server (sender)
// sending loop ...
// file sent
shutdown(socket, SHUT_WR);
// wait acknowledgement from the client
ack = read(socket, buffer, bufsize);
if (ack < 0) {
// handle errors
}
Client (receiver)
// receiving loop..
if (filesize > 0) {
// incomplete transfer
// handle error
}
// send acknowledgement to the server
// this will send a FIN and trigger a read = 0 on the server
shutdown(socket, SHUT_WR);
close(socket);
What about tricks number 2, 3 and 4?
Trick number 2 is not needed because as soon as the server accepts the connection the application proceed to the file transfer. NO extra messages are exchanged.
Trick number 3 is already implemented
Trick number 4 is also already implemented. As mentioned earlier the file size is hardcoded, so there is no need to exchange it.
Did this solve my original problem?
NO my problem was not solved. The error is still happening, and as of today, it only happened when testing the application with both client and server on localhost.
What do you think?
Is there a way to prevent this?
You're:
assuming that read fills the buffer, even though
you're defending magnificently against write() not writing the entire buffer.
You need to do (1), and you don't need to do (2) because you're in blocking mode and POSIX assures that write() doesn't return until all the data is written.
A simple version of both loops:
while ((nread = read(inFD, buffer, 0, sizeof buffer)) > 0)
{
write(outFD, buffer, 0, nread);
}
if (nread == -1)
; // error
A more correct version would check the result of write() for errors of course.

Unable to receive the last socket data sent using the MSG_MORE flag

server side code:
dirp=opendir(path);
if(dirp==NULL)
{
strcpy(err,"error:");
strcat(err,strerror(errno));
send(fd,err,sizeof(err),0);
}
else
{
printf("\nstream opened\n");
while((dp=readdir(dirp))!= NULL)
{
r=send(fd,dp->d_name,100,MSG_MORE);
if(r<0)
perror("\nsend:");
printf("\n%s",dp->d_name);
}
}
client:
while(recv(mainsocket,lsbuf,100,0)>0)
{
printf("\n %s",lsbuf);
bzero(lsbuf,sizeof(lsbuf));
}
the server side is printing all the filenames on the standard output,but on the client side the client is not receiving the last filename and program is getting blocked at that point
The problem is with the send syscall. You call it with MSG_MORE flag that means the more data will follow and send waits for more data without actually sending. The last chunk of data you should send without this flag. Thus your server side should look like:
dp = readdir(dirp);
if (dp != NULL)
{
/* each time check whether there are more data */
while((dp_next = readdir(dirp))!= NULL)
{
r = send(fd, dp->d_name, 100, MSG_MORE);
if (r < 0) {
perror("\nsend");
}
printf("\n%s",dp->d_name);
dp = dp_next;
}
/* send the last or the only record */
r = send(fd, dp->d_name, 100, 0);
if (r < 0) {
perror("\nsend");
}
}
Another posibility to fix the problem is to close the conenction with the close(fd) syscall. It send all data in the buffer before closing the connection. It's a less clean, but more simple solution.
Your client prints the newline before lsbuf, hence everything since the previous newline is lost in your output buffer.
Four solutions:
use printf("%s\n", lsbuf); instead of ..."\n %s"...
use puts(lsbuf);, which has the same effect, but is slightly more appropriate
use fflush(stdout) after your client loop to flush the output buffer
use unbuffered output, see setvbuf() for details
Note that this problem doesn't seem to be networking-related (although I'd substitute MSG_MORE with 0), it's merely a problem with output buffering.
On a sidenote, I strongly suggest to send strlen(dp->d_name) + 1 bytes instead of 100 bytes. This way, you won't send more bytes than necessary, and on the other hand you won't truncate the output if your directory entries happen to be larger than 100 bytes.
Also, neither your client nor your server checks whether send()/recv() returns 0, which means that the connection has been closed by the remote end.

Resources