I have a client that need to read on a socket a sequence of char sent by server.
Client read with system-call read() a socket SOCK_DGRAM.
Here the complete function with system-call read(..) inside.
ssize_t readLine(int sockd, void *vptr, size_t maxlen)
{
ssize_t n, rc;
char c, *buffer;
buffer = vptr;
for ( n = 1; n < maxlen; n++ )
{
rc = read(sockd, &c, 1);
if ( rc == 1 )
{
*buffer++ = c;
if (c == '\0') break;
}
else
{
if (errno == EINTR) continue;
return -1;
}
}
*buffer = 0;
return n;
}
The problem is that if Server send a sequence of char like this ABCDEF'\0', this client read only A and then the system-call read() go in blocking mode.
I have used Wireshark to see if server work well and it send correctly ABCDEF'\0' in a UDP packet. All ok from this point of view.
Thanks to all in advance.
With datagram sockets, you need to read and write the whole datagram at once.
If you don't give read enough space to read the entire datagram, the rest of the datagram simply disappears.
int datagram_length = read(sockd, vptr, maxlen - 1);
if (datagram_length < 0) {
// complain about the error
} else {
vptr[datagram_length] = 0;
}
Related
I am trying to build a chat application between the server and the client. My doubt is for sending information from the client or from the server I was able to handle the partial send with the help of the loop, but I am unable to find out the length of the send data bytes from the client to the server or from the server to the client, thereby having problem in creating the memory for the received bytes and printing.
My chat function code for the client:
int chat_function(int sockfd)
{
char ch;
char *buf;
char *newp;
int ret_send = 0;
int ret_recv = 0;
int buf_size = 0;
while(1) {
printf("From client, enter the message : ");
buf = (char *)malloc(sizeof(char));
if (buf == NULL)
return -1;
while ((ch = getchar()) != '\n') {
buf[buf_size++] = ch;
newp = (char *)realloc(buf, (buf_size + 1) * sizeof(char));
if ( newp == NULL) {
free(buf);
return -1;
}
buf = newp;
}
buf[buf_size] = '\0';
ret_send = send_all(sockfd, buf, buf_size);
if (ret_send == -1)
error(1, errno, "error in send() function call\n");
memset(buf, 0, buf_size);
ret_recv = recv_all(sockfd, buf, buf_size);
if (ret_recv == -1) {
error(1, errno, "error in recv() function call\n");
} else if (ret_recv == -2) {
printf("Oops the server has closed the connection\n");
free(buf);
break;
}
printf("From Server : %s", buf);
if ((strncmp(buf, "exit", 4)) == 0) {
printf("Client Exit...\n");
free(buf);
break;
}
free(buf);
}
}
For handling partial send:
int send_all(int sockfd, char *buf, int buf_size)
{
int bytes_left = 0;
size_t send_bytes = 0;
bytes_left = buf_size
while (1) {
send_bytes = send(fd, buf, bytes_left, 0);
if (send_bytes == -1)
return -1;
buf = buf + send_bytes;
bytes_left = bytes_left - send_bytes;
if (bytes_left == 0)
break;
}
return 0;
}
TCP is a stream protocol, meaning there are no message boundaries: it is just a full-duplex (meaning data flows in both directions at the same time, as if there were two separate lanes) more or less continuous stream of data.
UDP is a datagram protocol, and does have message boundaries. There is an ioctl (FIONREAD/SIOCINQ) that provides the length of the next datagram, but because it involves a syscall, doing that for every message you receive is going to be slow and inefficient. Instead, you normally use a buffer large enough to hold the largest acceptable message, and copy it if/when necessary. However, UDP also has no reliability guarantees, and often UDP datagrams are completely lost without any trace or discernible reason; that's just what happens.
For a chat client-server connection, you'll want to use TCP.
Since the underlying connection is just a stream of data, you need to design a protocol for the communications, so that the stream can be split into messages, with each message processed separately.
The simplest case would be to use the nul character, \0, as a message separator.
The "send" function would then look something like this:
/* Returns 0 if message successfully sent,
nonzero errno code otherwise. */
int send_message(int descriptor, const char *message)
{
/* If message is NULL, we cannot use strlen(); use zero for that. */
const size_t message_len = (message) ? strlen(message) : 0;
/* Temporary variables for the sending part. */
const char *ptr = message;
const char *const end = message + message_len + 1; /* Include '\0' at end */
ssize_t bytes;
/* Check valid descriptor and message length. */
if (descriptor == -1 || message_len < 1)
return errno = EINVAL;
/* Write loop for sending the entire message. */
while (ptr < end) {
bytes = write(descriptor, ptr, (size_t)(end - ptr));
if (bytes > 0) {
ptr += bytes;
} else
if (bytes != -1) {
/* This should never happen. */
return errno = EIO;
} else
if (errno != EINTR) {
/* We do not consider EINTR an actual error; others we do. */
return errno;
}
}
return 0;
}
The above send_message() function writes the specified string, including the string terminating nul character \0, to the specified descriptor.
On the read end, we need a buffer large enough to hold at least one full message. Instead of always waiting for incoming data, we need to check if the buffer already contains a full message, and if it does, return that. Also, you do not necessarily want to always wait for an incoming message, because that would mean you cannot send two messages in a row.
So, here's my suggestion:
static int incoming_desc = -1;
static char *incoming_data = NULL;
static size_t incoming_size = 0;
static char *incoming_next = NULL; /* First received but not handled */
static char *incoming_ends = NULL; /* Last received but not handled */
#define INCOMING_CHUNK 4096
/* Receive a new message into dynamically allocated buffer,
and return the length. Returns 0 when no message, with errno set.
Waits at most ms milliseconds for a new message to arrive.
errno == EAGAIN: no message, timeout elapsed.
errno == ECONNABORTED: other end closed the connection.
*/
size_t get_message(char **message, size_t *size, long ms)
{
struct timeval timeout;
/* Make sure the parameters are sane. */
if (!message || !size || ms < 0) {
errno = EINVAL;
return 0;
}
/* For this function to work like getline() and getdelim() do,
we need to treat *message as NULL if *size == 0. */
if (!*size)
*message = NULL;
timeout.tv_sec = ms / 1000;
timeout.tv_usec = (ms % 1000) * 1000;
/* Timeout loop. */
while (1) {
fd_set readfds;
ssize_t bytes;
size_t used;
int result;
/* Is there a pending complete message in the buffer? */
if (incoming_ends > incoming_next) {
char *endmark = memchr(incoming_next, '\0', (size_t)(incoming_ends - incoming_next));
if (endmark) {
const size_t len = (size_t)(endmark - incoming_next) + 1;
/* Reallocate the message buffer, if necessary. */
if (len > *size) {
char *temp = realloc(*message, len);
if (!temp) {
errno = ENOMEM;
return 0;
}
*message = temp;
*size = len;
}
/* Copy message, */
memcpy(*message, incoming_next, len);
/* and remove it from the buffer. */
incoming_next += len;
/* In case the other end sent just the separator, clear errno. */
errno = 0;
/* We return the length sans the separator. */
return len - 1;
}
}
/* Do we have time left to check for input? */
if (timeout.tv_sec <= 0 && timeout.tv_usec <= 0)
break; /* Nope. */
/* Is incoming_desc one we can select() for? */
if (incoming_desc < 0 || incoming_desc >= FD_SETSIZE)
break; /* Nope. */
FD_ZERO(&readfds);
FD_SET(incoming_desc, &readfds);
result = select(incoming_desc + 1, &readfds, NULL, NULL, &timeout);
if (result < 1)
break; /* Nothing interesting happened (we ignore error here). */
if (!FD_ISSET(incoming_fd, &readfds))
break;
/* Number of bytes used in the buffer right now. */
used = (size_t)(incoming_ends - incoming_data);
/* Do we have at least INCOMING_CHUNK bytes available? */
if (used + INCOMING_CHUNK >= incoming_size) {
/* Nope. Repack the incoming buffer first. */
if (incoming_next > incoming_data) {
const size_t len = (size_t)(incoming_ends - incoming_next);
if (len > 0)
memmove(incoming_data, incoming_next, len);
incoming_next = incoming_data;
incoming_ends = incoming_data + len;
}
/* Recalculate the number of bytes we have free now. Enough? */
used = (size_t)(incoming_ends - incoming_data);
if (used + INCOMING_CHUNK > incoming_size) {
/* Grow incoming buffer. */
const size_t newsize = used + INCOMING_CHUNK;
char *temp = realloc(incoming_data, newsize);
if (!temp) {
errno = ENOMEM;
return 0;
}
incoming_next = temp + (size_t)(incoming_next - incoming_data);
incoming_ends = temp + used;
incoming_data = temp;
incoming_size = newsize;
}
}
/* Read more data into the buffer; up to a full buffer. */
bytes = read(incoming_fd, incoming_ends, incoming_size - used);
if (bytes > 0) {
incoming_ends += bytes;
} else
if (bytes == 0) {
/* Other end closed the connection. We may have a partial message
in the buffer, and should handle that too, but for now, we
just error out. */
errno = ECONNABORTED;
return 0;
} else
if (bytes != -1) {
/* Should never happen. */
errno = EIO;
return 0;
} else
if (errno == EINTR || errno == EAGAIN || errno == EWOULDBLOCK) {
/* No data yet, interrupted by signal delivery, etc. */
continue;
} else {
/* errno is set to indicate which error happened. */
return 0;
}
}
/* Timeout. */
errno = EAGAIN;
return 0;
}
Note that get_message() works like getline(): you do e.g.
char *msg = NULL;
size_t size = 0;
size_t len;
len = get_message(&msg, &size, 100); /* 100 ms = 0.1 seconds */
if (len) {
/* msg contains a full message of len characters */
} else
if (errno == ECONNABORTED) {
/* Other end closed the connection */
} else
if (errno != EAGAIN) {
fprintf(stderr, "Error receiving data: %s.\n", strerror(errno));
}
Then, you can reuse the same dynamically allocated buffer by just calling e.g.
len = get_message(&msg, &size, 100); /* 100 ms = 0.1 seconds */
again.
There is no such mechanism built into TCP or UDP. You need to implement your own protocol on top of it. One of the possible solutions is:
If the content delivered is static.
If the sending end knows the size of the data that is being delivered prior, your client and server can agree on specific terms. For example, the first four bytes sent by the server is the size of the remaining message represented in network byte order.
Server code
uint32_t n_size = htonl(size); // Convert the data size into network byte order.
write(sockfd, &n_size, sizeof(n_size)); // Send to the client.
Client code
uint32_t n_size;
int n_read = 0;
for ( ; ; ) {
int rd_status = read(sockfd, (void*) &n_size + n_read, sizeof(n_size) - n_read);
if (rd_status <= 0)
goto handle_this_case;
n_read = n_read + rd_status;
if (n_read == sizeof(n_size))
break;
}
uint32_t size = ntohl(n_size);
If the content delivered is generated on the fly.
In this case, even the server is not aware of the size of the message. You need to build your functions for handling this case. Below I have shown a bare minimal implementation:
Client-Side:
struct data_unit
{
void* data;
int size;
};
struct data_storage
{
struct data_unit unit;
struct data_storage* next;
};
void append_data(struct data_storage* storage, struct data_unit* unit);
struct data_unit* dump_data(struct data_storage* storage);
int main()
{
struct data_storage storage;
struct data_unit unit;
unit.data = malloc(MAX_SIZE);
for ( ; ; ) {
int rd_status = read(sockfd, unit.data, MAX_SIZE);
if (rd_status < 0)
goto handle_this_case;
else if (rd_status == 0)
break;
unit.size = rd_status;
append_data(&storage, &unit);
}
struct data_unit* t_data = dump_data(&storage);
}
I have a checkpoint file that receives a server state. This states represents serialized commands that pass trough my network.
I'm trying to read the file but the it gets stuck on the read while loop.
My read function:
struct message_t *pmanager_readop(int fd){
if (fd < 0) return NULL;
// Variables
char *buffer = NULL;
int result, msg_size;
struct message_t *msg;
// Check if file has data
lseek (fd, 0, SEEK_END);
int size_ckp = lseek(fd, 0, SEEK_CUR);
if (size_ckp <= 0)
return NULL;
// Read message size
result = read_all(fd, (char *) &msg_size, 4);
if (result < 0) {
return NULL;
}
msg_size = ntohl(msg_size);
// ............
My read_all() function:
int read_all(int sock, char *buf, int len){
int bufsize = len;
while(len > 0){
int res = read(sock,buf,len);
if(res < 0){
if (errno == EINTR) continue;
return res;
}
buf += res;
len -= res;
}
return bufsize;
}
I use this same function to read data from my server/client connection with the same serialization and format but with a socket descriptor, and it works perfectly.
You ought to handle the case that read() returns 0, telling you that the other side shut-down the connection if reading from a socket descriptor, or EOF encountered if reading from a file descriptor.
Got some trouble with TCP socket multiplexing.
//socket is non-blocking
const int MAX = 4096;
char *buff[MAX];
char *p = buff;
int fd, rvalue;
rvalue = 0;
if ( (fd = open(path, O_RDONLY)) < 0 ) {
return errno;
} else {
int didsend, didread;
int shouldsend;
while ((didread = read(fd, buff, MAX)) > 0) {
p = buff;
shouldsend = didread;
while ( 1 ) {
didsend = send(sockfd, p, shouldsend, 0);
//if send succeeds and returns the number of bytes fewer than asked for then try to send rest part in next time.
if (didsend < shouldsend) {
p += didsent;
shouldsend -= didsend;
continue;
}
//if there is no place for new data to send, then wait a brief time and try again.
if ( didsend < 0 && (errno == EWOULDBLOCK || errno == EAGAIN) ) {
usleep(1000);
continue;
}
//if all data has been sent then sending loop is over.
if (didsend == shouldsend) {
break;
}
//send error
if ( didsend < 0 ) {
rvalue = errno;
break;
}
}
}
close(fd);
if (didread == -1) {
return errno;
}
return rvalue;
}
Assume I use an I/O Multiplexing function poll() or kqueue(), and non-blocking socket, then if there are only some small data like send a short message, it works fine.
But if it comes to large data, I mean larger than send()'s buffer size, since using non-blocking socket, send() will just send a portion of data, and return how much data it sends, the rest part of data can only be sent in another call of send(), but it takes time, and can't tell how long it will takes. So the second while() is actually a blocking send which using non-blocking socket.
Equivalent to：
//socket is blocking
const int MAX = 4096;
char *buff[MAX];
int fd, n;
if ( (fd = open(path, O_RDONLY)) < 0 ) {
return errno;
} else {
while ((n = read(fd, buff, MAX)) > 0) {
if (send(sockfd, buff, n, 0) < 0) {
return errno;
}
}
close(fd);
return 0;
}
So, what is the solution to this, multithreading might work but that's kind of wasting resource maybe.
This is the general pattern for a single-threaded server that works with multiple connections and non-blocking sockets.
It's primarily pseudo-code in C and doesn't do the necessary error checking. But it gives you an idea that for each accepted connection, you keep a struct instance that maintains the socket handle, request parsing state, response stream, and any other "state" members of that connection. Then you just loop using "select" to wait or having multiple threads doing this same thing.
Again this is only pseudo-code and uses select/poll as an example. You can get even more scalability with epoll.
while (1)
{
fd_set readset = {};
fd_set writeset = {};
for (int i = 0; i < number_of_client_connections; i++)
{
if (client_connections[i].reading_request)
FD_SET(client_connection.sock, &readset);
else
FD_SET(client_connection.sock, &writeset);
}
// add the listen socket to the read set
FD_SET(listen_socket, &readset);
select(n + 1, &readset, &writeset, &timeout); // wait for a socket to be ready (not shown - check for errors and return value)
if (FD_ISSET(listen_socket, &readset))
{
int new_client_socket = accept(listen_socket, &addr, &addrlength);
// create a struct that keeps track of the connection state data
struct ConnectionData client_connection = {};
client_connection.sock = new_client_socket;
client_connection.reading_request = 1; // awaiting for all the request bytes to come in
client_connections[number_of_client_connections++] = client_connection; // pseudo code, add the client_connection to the list
}
for (int i = 0; i < number_of_client_connections; i++)
{
if (client_connections[i].reading_request)
{
if (FD_ISSET(client_connections[i], &readset))
{
char buffer[2000];
int len = recv(client_connections[i].sock, buffer, 2000, 0);
// not shown - handle error case when (recv < 0)
// not shown - handle case when (recv == 0)
ProcessIncomingData(client_connections[i], buffer, len); // do all the request parsing here. Flip the client_connections[i].reading_request to 0 if ready to respond
}
}
else if (client_connections[i].reading_request == 0)
{
if (FD_ISSET(client_connections[i], &writeset))
{
client_connection* conn = &client_connections[i];
int len = send(conn->sock, conn->response_buffer + conn->txCount, conn->response_size - conn->txCount, 0);
conn->txCount += len;
if (conn->txCount == conn->response_size)
{
// done sending response - we can close this connection or change it to back to the reading state
}
}
}
}
I'm trying to implement a working HTTP Client-Server application just to make practice with network programming.
The 2 programs have to follow this basic algorithm:
CLIENT - send a GET request
SERVER - send "+OK\r\n"
SERVER - send file size in bytes
SERVER - send file
CLIENT - send ACK
I'm having a lot of troubles in the reading part, probably because i perform some dirty read on the stream.
These are the 2 reading function that i'm using:
/* Reads a line from stream socket s to buffer ptr
The line is stored in ptr including the final '\n'
At most maxlen chasracters are read*/
int readline (SOCKET s, char *ptr, size_t maxlen)
{
size_t n;
ssize_t nread;
char c;
for (n=1; n<maxlen; n++)
{
nread=recv(s, &c, 1, 0);
if (nread == 1)
{
*ptr++ = c;
if (c == '\n')
break;
}
else if (nread == 0) /* connection closed by party */
{
*ptr = 0;
return (n-1);
}
else /* error */
return (-1);
}
*ptr = 0;
return (n);
}
and:
int readNumber(SOCKET s, long *num, int maxRead)
{
size_t n;
ssize_t nread;
int totRead;
long number=0;
for (n=1; n<maxRead+1; n++)
{
nread=recv(s, &number, sizeof(number), 0);
if (nread == sizeof(number))
{
totRead+=nread;
*num = number;
}
else if (nread == 0) /* connection closed by party */
{
*num = 0;
return (n-1);
}
else /* error */
{
printf("nread = %d\n", nread);
return (-1);
}
}
return (totRead);
}
this is the snippet of the main where i receive the +OK message and then the file size:
memset(rbuf,0,sizeof(rbuf)); //rbuf is the buffer where is store the read
printf("waiting for response...\n");
result = readline(s, rbuf, sizeof(rbuf)); //reading function is above
printf("Byte read(okMsg) = %d\n", result);
if (result <= 0)
//ERROR MANAGEMENT
{
printf("Read error/Connection closed\n");
closesocket(s);
SockCleanup();
exit(1);
}
else
{
long fileLength=0;
unsigned char *fBuf;
//RECEIVE OK
if(!strcmp(rbuf,"+OK\r\n"))
{
puts("+OK\n");
//RECEIVE FILE LEN
int nw = readNumber(s, &fileLength, 1); //reading function is above
printf("Byte read(fDim) = %d\n", nw);
printf("File is %ld bytes long\n", fileLength);
if(nw >0)
{
// RECEIVE FILE
}
}
}
When i send the "+OK\r\n" string the server tells me that it sends 8 bytes, but when i read i find the '\0' char only after 6 bytes.
By the way it reads correctly the message, but when i try to read the file size (that is a long) it gives me back a wrong number.
My opinion is that the stream buffer is dirty, and that i'm reading 2 bytes that are not part of the file size, but i'm not understanding why this happens.
Please ask me more info if i'm not clear enough.
SOLVED:
Thank you all for your answers!!!
You put me in the right mindset to understand what was wrong.
Look like the problem was this declaration in the server:
char *okMsg = "+OK\r\n";
instead of
char okMsg[] = "+OK\r\n";
that lead me to an undefined behavior.
long number=0;
for (n=1; n<maxRead+1; n++)
{
nread=recv(s, &number, sizeof(number), 0);
You forgot to design and implement a protocol to carry the data between your server and your client. Because TCP provides a stream of bytes, your protocol should be defined as a stream of bytes.
How many bytes convey this number? Is "however many bytes a 'long' happens to occupy on my platform" a good answer? What's the semantic meaning of the first byte? Is "whatever the first byte of a 'long' happens to mean on my platform" a good answer?
A good answer would be, "The size shall be conveyed as a 4-byte unsigned integer in little-endian byte order". Then make absolutely sure your code sends and receives in that format.
I have 2 machines running a simple C TCP server that I have written for testing purposes, 1 with Fedora 16, the other with Ubuntu 11.10. My Fedora machine works perfectly but on the Ubuntu machine, recv() does not block. Please keep in mind that these machines are running the same exact code. Has anybody seen this before? Thanks
int TcpSocket::ReadFromClient(int socket, char* buf, int len)
{
char *request = buf;
int slen = len;
int c = recv(socket, request, slen, 0);
while((c > 0) && (request[c-1] != '\n'))
{
request += c;
slen -= c;
c = recv(socket, request, slen, 0);
}
if (c < 0)
{
return c;
}
else if(c == 0)
{
//Sending back an empty string
buf[0] = '\0';
}
return len-slen;
}
It looks like the intention of your code is to stop reading when a '\n' byte arrives. If that is the case then you need to read from the socket 1 byte at a time instead of using the entire available buffer size, especially since you are only checking the last byte of the buffer instead of checking every byte received.
You should also change the loop logic to only call recv() in one place instead of two places. Your current implementation is calling recv() with slen=0 when the buffer is exhausted, which will set c=0 and nullify the first byte in the buffer.
Try this instead:
int TcpSocket::ReadFromClient(int socket, char* buf, int len)
{
int slen = len;
char ch;
while (len > 0)
{
int ret = recv(socket, &ch, 1, 0);
if (ret > 0)
{
*buf = ch;
++buf;
--len;
if (ch == '\n')
break;
}
else
{
if ((ret == 0) || (errno != EAGAIN))
return ret;
fd_set readfd;
FD_ZERO(&readfd);
FD_SET(socket, &readfd);
timeval tv;
tv.tv_sec = 5;
tv.tv_usec = 0;
ret = select(socket+1, &readfd, NULL, NULL, &tv);
if (ret < 0)
return ret;
if (ret == 0)
{
// timeout elapsed while waiting for data
// do something if desired...
}
}
}
return slen - len;
}