I'm writing an embedded Linux application that (1) opens a serial connection to another device, (2) sends a known command, (3) checks port for incoming characters (response) until expected response phrase or character is detected, (4) repeats step 2 and 3 until a series of commands are sent and responses received, (5) then closes the port.
My app would go through some cycles of the above sequence and every now and then it would be waiting for a response (reading) when all of a sudden the communication stops and my software faults out because of my built-in timeout logic.
Is there anything in my port configuration that would cause the port to be blocked due to specific byte sent (possibly due to electrical noise)?
Here is how I'm opening my ports (showing configurations via termios.h):
struct termios options;
fd = open("/dev/ttyUSB0", O_RDWR | O_NOCTTY | O_NONBLOCK);
if (fd == -1) {
debug() << "Port open failed!"
return FAIL;
}
debug() << "Port Opened Successful"
fcntl(fd, F_SETFL, 0); // This setting interacts with VMIN and VTIME below
// Get options
tcgetattr(fd, &options);
// Adjust Com port options
options.c_cflag |= (CLOCAL | CREAD); // Program will not "own" port, enable reading on port
options.c_lflag &= ~(ICANON | ECHO | ECHOE | ISIG); // Sets RAW input mode (does not treat input as a line of text with CR/LF ending)
options.c_oflag &= ~ OPOST; // Sets RAW ouput mode (avoids newline mapping to CR+LF characters)
options.c_iflag &= ~(IXON | IXOFF | IXANY); // Turns off SW flow c
options.c_cc[VMIN] = 0;
options.c_cc[VTIME] = 10;
// Set options
tcsetattr(fd, TCSANOW, &options);
//return fd;
return SUCCEED;
I can't figure out why the communication all of a sudden just freezes up and then goes away when I cycle power to my device. Thanks all!
More info - here are my read and write functions:
int Comm::Receive(unsigned char* rBuf)
{
int bytes;
ioctl(fd, FIONREAD, &bytes);
if (bytes >= 1)
{
bytes = read(fd, rBuf, 1);
if (bytes < 0)
return READ_ERR;
return SUCCEED;
}
else
return NO_DATA_AVAILABLE;
}
int Comm::Send(int xCt, unsigned char* xBuf)
{
int bytes;
if (fd == -1)
return FAIL;
bytes = write(fd, xBuf, xCt);
if (bytes != xCt)
return FAIL;
else
return SUCCEED;
}
Welcome to the joys of serial ports...
Thought 1: wrap your read calls with a select ()
Thought 2: Unset the ICANON flag in tcsetattr, and set a VTIME attribute for a deliberate timeout (and then, obviously, handle it)
Thought 3: Nothing about serial comms ever works perfectly.
I also had a similar problem with sending command to devices and reading responses from the device. Please refer to below SOF post and the answer this was working for my problem.
In these cases, We have to care about the protocol which we are going to use for device communication (send and receive). If we can send commands successfully and we didn't receive a response with a noise from device implies there is something wrong in the data packet which was sent to the device. First of all check the protocol specification first and create a byte array for a simple command (like make a beep sound) and send it.
Send data to a barcode scanner over RS232 serial port
I can do something for you If you can post your complete source code with the output.
Enjoy the code. Thanks.
Related
I am testing a serial communication protocol based on printable characters only.
The setup has a pc connected to an arduino board by USB. The PC USB serial is operated in canonical mode, with no echo, no flow control, 9600 baud.
Since a read timeout is requested, pselect is called before the serial read. The arduino board simply echoes back every received character without any processing. The PC OS is Linux Neon with kernel 5.13.0-40-generic.
When lines of a specific length are transmitted from the PC and echoed back by the arduino, they are received correctly except for the final new line that is missing.
A further read, returns an empty line (the previously missing NL).
Lines with different length are transmitted and received correctly, including the trailing NL.
This behavior is fully repeatable and stable. The following code reproduce the problem for a line transmitted with a length of 65 characters (including NL) and received with a length of 64 (NL missing). Other line lengths work fine.
Thanks for any hints.
/* remote serial loop test 20220626 */
#include <errno.h>
#include <fcntl.h>
#include <stdio.h>
#include <stdint.h>
#include <stdlib.h>
#include <sys/select.h>
#include <string.h>
#include <termios.h>
#include <time.h>
#include <unistd.h>
#define TX_MINLEN 63
#define TX_MAXLEN 66
#define DATA_MAXLEN 128
#define LINK_DEVICE "/dev/ttyUSB0"
#define LINK_SPEED B9600
#define RECEIVE_TIMEOUT 2000
int main()
{
int wlen;
int retval;
int msglen;
uint8_t tx_data[257] = {'0','1','2','3','4','5','6','7','8','9','A','B','C','D','E','F'};
for (int i=16; i < 256; i++) tx_data[i] = tx_data[i & 0xf];
uint8_t rx_data[257];
/* serial interface */
char * device;
int speed;
int fd;
fd_set fdset;
struct timespec receive_timeout;
struct timespec *p_receive_timeout = &receive_timeout;
struct termios tty;
/* open serial device in blocking mode */
fd = open(LINK_DEVICE, O_RDWR | O_NOCTTY);
if (fd < 0) {
printf("Error opening %s: %s\n",LINK_DEVICE,strerror(errno));
return -1;
}
/* prepare serial read by select to have read timeout */
FD_ZERO(&(fdset));
FD_SET(fd,&(fdset));
if (RECEIVE_TIMEOUT >= 0) {
p_receive_timeout->tv_sec = RECEIVE_TIMEOUT / 1000;
p_receive_timeout->tv_nsec = RECEIVE_TIMEOUT % 1000 * 1000000;
}
else
p_receive_timeout = NULL;
/* get termios structure */
if (tcgetattr(fd, &tty) < 0) {
printf("Error from tcgetattr: %s\n", strerror(errno));
return -1;
}
/* set tx and rx baudrate */
cfsetospeed(&tty, (speed_t)LINK_SPEED);
cfsetispeed(&tty, (speed_t)LINK_SPEED);
/* set no modem ctrl, 8 bit, no parity, 1 stop */
tty.c_cflag |= (CLOCAL | CREAD); /* ignore modem controls */
tty.c_cflag &= ~CSIZE;
tty.c_cflag |= CS8; /* 8-bit characters */
tty.c_cflag &= ~PARENB; /* no parity bit */
tty.c_cflag &= ~CSTOPB; /* only need 1 stop bit */
tty.c_cflag &= ~CRTSCTS; /* no hardware flowcontrol */
/* canonical mode: one line at a time (\n is line terminator) */
tty.c_lflag |= ICANON | ISIG;
tty.c_lflag &= ~(ECHO | ECHOE | ECHONL | IEXTEN);
/* input control */
tty.c_iflag &= ~IGNCR; /* preserve carriage return */
tty.c_iflag &= ~INPCK; /* no parity checking */
tty.c_iflag &= ~INLCR; /* no NL to CR traslation */
tty.c_iflag &= ~ICRNL; /* no CR to NL traslation */
tty.c_iflag &= ~IUCLC; /* no upper to lower case mapping */
tty.c_iflag &= ~IMAXBEL;/* no ring bell at rx buffer full */
tty.c_iflag &= ~(IXON | IXOFF | IXANY);/* no SW flowcontrol */
/* no output remapping, no char dependent delays */
tty.c_oflag = 0;
/* no additional EOL chars, confirm EOF to be 0x04 */
tty.c_cc[VEOL] = 0x00;
tty.c_cc[VEOL2] = 0x00;
tty.c_cc[VEOF] = 0x04;
/* set changed attributes really */
if (tcsetattr(fd, TCSANOW, &tty) != 0) {
printf("Error from tcsetattr: %s\n", strerror(errno));
return -1;
}
/* wait for serial link hardware to settle, required by arduino reset
* triggered by serial control lines */
sleep(2);
/* empty serial buffers, both tx and rx */
tcflush(fd,TCIOFLUSH);
/* repeat transmit and receive, each time reducing data length by 1 char */
for (int l=TX_MAXLEN; l > TX_MINLEN - 1; l--) {
/* prepare data: set EOL and null terminator for current length */
tx_data[l] = '\n';
tx_data[l+1] = 0;
/* send data */
int sent = write(fd,tx_data,l+1);
/* receive data */
/* wait for received data or for timeout */
retval = pselect(fd+1,&(fdset),NULL,NULL,p_receive_timeout,NULL);
/* check for error or timeout */
if (retval < 0)
printf("pselect error: %d - %s\n",retval,strerror(errno));
else if (retval == 0)
printf("serial read timeout\n");
/* there is enough data for a non block read: do read */
msglen = read(fd,&rx_data,DATA_MAXLEN);
/* check rx data length */
if (msglen != l+1)
printf("******** RX ERROR: sent %d, received %d\n",l+1,msglen);
else
continue;
/* check received data, including new line if present */
for (int i=0; i < msglen; i++) {
if (tx_data[i] == rx_data[i])
continue;
else {
printf("different rx data:|%s|\n",rx_data);
break;
}
}
/* clear RX buffer */
for (int i=0; i < msglen + 1; i++) rx_data[i] = 0;
}
}
When performing stream-based communication, be it via pipes, serial ports, or TCP sockets, you can never rely on reads to always return a full "unit of transmission" (in this case a line, but could also be a fixed-size block). The reason is that stream-based communication can always be split by any part of the transmission stack (even potentially the sender) into multiple blocks, and there is never a guarantee that a single read will always read a full block.
For example, you could always run into the race condition that your microcontroller is still sending parts of the message when you call read(), so not all characters are read exactly. In your case, that's not what you're seeing, because that would be more of a stochastic phenomenon (that would be worse with an increased interrupt load on the computer) and not so easily reproducible. Instead, because you're talking about the number 64 here, you're running into the static buffer size used in the kernel's tty driver that will only ever return at most 64 bytes at once, regardless of what the specified read size actually is. However, in other cases it could still be that you'll see additional failures by the kernel returning only the first couple of characters of a line in the first read(), and the rest in the second, depending on precise timing details -- you've probably not seen that yet, but it's bound to happen at some point.
The only reliable way to properly implement communication protocols in streaming situations (serial port, pipes, TCP sockets, etc.) is to consider the following:
For fixed-size data (e.g. communication units that are always N bytes in size) to loop around a read() call until you've read exactly the right amount of bytes (reads that follow an incomplete read would obviously ask for less bytes than the original read, just to make up the difference)
For variable-size data (for example communication units that are separated by a line end character) you have two options: either you read only one character at a time until you reach the end-of-line character (inefficient, uses lots of syscalls), or you keep track of the communication state via a large enough buffer that you constantly fill with read() operations until the buffer contains a line-end character, at which point you remove that line from the buffer (but keep the rest) and process that.
As a complete aside, if you're doing anything with serial communication, I can very much recommend the excellent libserialport library (LGPLv3 license) that makes working with serial ports a lot easier -- and has the benefit of being cross-platform. (Doesn't help with your issue, just thought that I'd mention it.)
Upgrading from linux kernel version 5.13.0-40-generic to 5.13.0-51-generic solved the problem.
I am using a UART on a single board computer (Olimex A13) and I am trying to send and receive data through the UART. On the send side I have no problems. The Olimex board sends data to the serial buss and it is received on an external PC that is running putty terminal.
When I try to go in the other direction (PC sends data to be received by the Olimex board) I am running into a weird problem. When I start to send data, initially no data shows up on the Olimex board (I have tried 2 methods... cat /dev/ttyS0 and setting up a receive C program. Both exhibit the same results).
If I send data for a while (seems to vary from less than 10 characters to something over 20) at some point the uart seems to "wake up" and now data appears on the Olimex board. Once the data transfer starts it will keep going as long as I want and at fast or slow data rates. If I break out of the receive program or end the cat /dev/ttyS0 process and try to start receiving again, I need to go through the same process of "kicking" the uart until it decides it wants to receive again.
I have put an oscilloscope on the RX pin leading into the UART. The data coming in is at good levels and the data appears here from the very first send so the data is not being blocked initially and then arriving at the RX pin at some time later. So I am currently assuming that the issue is with a UART setting of some kind. Does anyone know of a UART setting that could possibly explain this behavior?
Below is the receive program that I am running. I don't think that there is an issue with it because I see the same behavior when I do a cat /dev/ttyS0. Any help here would be appreciated.
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <termios.h>
#include <stdio.h>
#define BAUDRATE B115200
#define MODEMDEVICE "/dev/ttyS0"
#define _POSIX_SOURCE 1 /* POSIX compliant source */
#define FALSE 0
#define TRUE 1
volatile int STOP=FALSE;
main()
{
int fd,c, res;
struct termios oldtio,newtio;
char buf[255];
fd = open(MODEMDEVICE, O_RDWR | O_NOCTTY );
if (fd <0) {perror(MODEMDEVICE); return(-1); }
tcgetattr(fd,&oldtio); /* save current port settings */
bzero(&newtio, sizeof(newtio));
newtio.c_cflag = BAUDRATE | CRTSCTS | CS8 | CLOCAL | CREAD;
newtio.c_iflag = IGNPAR;
newtio.c_oflag = 0;
/* set input mode (non-canonical, no echo,...) */
newtio.c_lflag = 0;
newtio.c_cc[VTIME] = 0; /* inter-character timer unused */
newtio.c_cc[VMIN] = 1; /* blocking read until 5 chars received */
tcflush(fd, TCIFLUSH);
tcsetattr(fd,TCSANOW,&newtio);
while (STOP==FALSE) { /* loop for input */
res = read(fd,buf,255); /* returns after 5 chars have been input */
buf[res]=0; /* so we can printf... */
printf("String:%s - Character Count:%d\n", buf, res);
if (buf[0]=='z') STOP=TRUE;
}
tcsetattr(fd,TCSANOW,&oldtio);
}
Sawdust made these comments....
"sends data to the serial buss" -- RS232 is not a "buss". "When I
start to send data" -- Your writing style is confusing. What kind of
data is sent to the Olimex? For the cat /dev/ttyS0 test, how did you
set up the serial port, or was it in an unknown mode? bzero(&newtio,
sizeof(newtio)); -- That is a dangerous way to setup the termios
structure. newtio.c_cflag = BAUDRATE... That is not the proper way to
set the baud rate; you need to use the cfset[io]speed() functions. You
write of a "kick", yet you don't mention doing anything explicit
ok, RS232 is not a buss, I misspoke....
The data I am sending is a stream of characters that are typed on Putty terminal.
For the /dev/ttyS0 test the settings are the same as are set in the c program. I verified this using stty.
I copied this code from another post. I was not sure what bzero did. If you could expand on "dangerous" that would be great. Or better if you have a better way to do what bzero does?
I will update the way to set baud rate.. but send is working perfectly and I verified using stty that baudrate was indeed 115200.
As far as "kicking" I mean that I need to send a number of bytes of data to the uart before any appear on the console. Others have mentioned a FIFO that may be an issue here but the first characters that are sent are lost. For example if I were to send the string of characters abcdefg.... I could type up until k and then l would appear on the console, and then any future characters after l would also appear in the console. If I were to go through the same process again after starting the receive program (or cat /dev/ttyS0) the point where the receive starts appearing could be in a different place (i.e. at h or n).
from this I copied the example for serial port configuration:
tcgetattr (serialfd, &tty);
cfsetospeed(&tty,B115200);
cfsetispeed(&tty,B115200);
tty.c_cflag = (tty.c_cflag & ~CSIZE) | CS8;
tty.c_iflag &= ~IGNBRK;
tty.c_lflag = 0;
tty.c_oflag = 0;
tty.c_cc[VMIN] = 0;
tty.c_cc[VTIME] = 5;
tty.c_iflag &= ~(IXON | IXOFF | IXANY);
tty.c_cflag |= (CLOCAL | CREAD);
tty.c_cflag &= ~(PARENB | PARODD);
tty.c_cflag |= 0;
tty.c_cflag &= ~CSTOPB;
tty.c_cflag &= ~CRTSCTS;
My actual code is like this:
char buf[100];
write(serialfd, "PING", strlen("PING"));
fsync(serialfd);
while (1)
{
read(serialfd, buf, sizeof(buf));
printf("length: %d\n", strlen(buf));
}
In this case it is printing length: 6 infinitely without stopping. when I change tty.c_cc[VMIN] = 1 and tty.c_cc[VTIME] = 0 it doesn't read.(it blocks in read())
I'm using debian 6.0.5 with usb to serial converter. I open serial port like this:
serialfd = open("/dev/ttyUSB0", O_RDWR | O_NOCTTY | O_SYNC);
Look at your code
while (1)
{
read(serialfd, buf, sizeof(buf));
printf("length: %d\n", strlen(buf));
}
You have written a packet prior to this loop, then on the first iteration you read the available data which gets read into your buffer. You need to either memset your buffer to zeros each time OR zero terminate your buffer using the read byte count given in the return value of your call to read. You then loop infinitely, each time reading again - but subsequent reads will not copy any more data as there is none to read. However, your buffer remains unchanged by the call to read therefore and your printed output remains the same every iteration as the buffer remains the same every iteration.
As for the blocking aspect, you should read the following guide (which has been recommended on SO before and is very good as an introduction to serial port programming)
http://www.easysw.com/~mike/serial/serial.html
This section describes the behaviour you get when setting VMIN and VTIME to various values. In particular the last paragraph explains the blocking behaviour you see.
VMIN specifies the minimum number of characters to read. If it is set
to 0, then the VTIME value specifies the time to wait for every
character read. Note that this does not mean that a read call for N
bytes will wait for N characters to come in. Rather, the timeout will
apply to the first character and the read call will return the number
of characters immediately available (up to the number you request).
If VMIN is non-zero, VTIME specifies the time to wait for the first
character read. If a character is read within the time given, any read
will block (wait) until all VMIN characters are read. That is, once
the first character is read, the serial interface driver expects to
receive an entire packet of characters (VMIN bytes total). If no
character is read within the time allowed, then the call to read
returns 0. This method allows you to tell the serial driver you need
exactly N bytes and any read call will return 0 or N bytes. However,
the timeout only applies to the first character read, so if for some
reason the driver misses one character inside the N byte packet then
the read call could block forever waiting for additional input
characters.
VTIME specifies the amount of time to wait for incoming characters in
tenths of seconds. If VTIME is set to 0 (the default), reads will
block (wait) indefinitely unless the NDELAY option is set on the port
with open or fcntl.
I've got this strange issue with reading from a serial port on Linux. About 50% of the time, I won't get any data back from the serial port when I'm reading from it, yet if I use GTKTerm I get data back 100% of the time. I've double-checked my settings, and the terminal settings that I'm using are the exact same that GTKTerm uses, so I'm mystified as to why this would occur. The only thing that I could think of is that the port is set in canonical mode, not raw mode, but I set it properly - the device that I'm reading from does not send back newlines with all of the commands. However, everything is working properly when the device sends back commands with a newline character.
Here's the code I'm using:
int fd = open(serial_port, O_RDWR );
if( fd < 0 ){
perror("open");
}
//Set the serial port settings
struct termios newio;
if( tcgetattr( fd, &newio ) < 0 )
perror("tcgetattr");
if( cfsetospeed( &newio, B9600 ) < 0 )
perror("cfsetospeed");
newio.c_cflag &= ~CSTOPB;
newio.c_cflag &= ~CSIZE;
newio.c_cflag |= CS8;
newio.c_cflag |= CREAD;
newio.c_iflag |= IGNPAR;
newio.c_iflag |= IGNBRK;
newio.c_iflag &= ~BRKINT;
newio.c_iflag &= ~ICRNL;
newio.c_iflag &= ~IXON;
newio.c_cflag |= CLOCAL;
newio.c_oflag = 0;
newio.c_lflag = 0;
newio.c_cc[VTIME] = 0;
newio.c_cc[VMIN] = 1;
if( tcsetattr( fd, TCSANOW, &newio ) < 0 )
perror("tcsetattr");
tcflush( fd, TCOFLUSH );
tcflush( fd, TCIFLUSH );
Reading code (in separate thread)
void* thr(void* ign){
char buffer[10];
while( 1 ){
int got = read(fd, buffer, 10);
buffer[got] = 0;
printf("got %s\n", buffer);
}
}
Well, I found out the problem.
For some bizzare reason, the device that I'm sending to doesn't seem to have a big buffer or something; it turns out that when I send commands as a string all at once, I won't always get a response back. Sleeping between sending characters fixes the problem(though I admit, it's not a very good solution). What was probably happening was that I was getting interrupted at points, which would allow enough time for the device to process some of its input. GTKTerm will send out each character as it is received, and it's impossible to type fast enough for this error to occur.
I have a system where I am seeing strange behavior with the serial ports that I don't expect. I've previously seen this on occasion with usb-to-serial adapters, but now I'm seeing it on native serial ports as well, with much greater frequency.
The system is set up to run automated tests and will first perform some tasks that cause a large amount of data to be outputted from the serial device while I do not have the ports open. The device will also reset itself. Only the tx/rx lines are connected. There is no flow control.
After these tasks complete, the testware opens the serial ports and immediately fails because it gets unexpected responses. When I reproduce this, I found that if I open the serial port in a terminal program, I see several kilobytes of old data (that appears to have been sent when the port was closed) immediately flushed out. Once I close this program, I can then run the tests as expected.
What could cause this to happen? How does Linux handle buffering the serial port when the device is closed? If I opened a device, made it send output, and then closed it without reading from it, would this cause the same problem?
The Linux terminal driver buffers input even if it is not opened. This can be a useful feature, especially if the speed/parity/etc. are set appropriately.
To replicate the behavior of lesser operating systems, read all pending input from the port as soon as it is open:
...
int fd = open ("/dev/ttyS0", O_RDWR | O_NOCTTY | O_SYNC);
if (fd < 0)
exit (1);
set_blocking (fd, 0); // disable reads blocked when no input ready
char buf [10000];
int n;
do {
n = read (fd, buf, sizeof buf);
} while (n > 0);
set_blocking (fd, 1); // enable read blocking (if desired)
... // now there is no pending input
void set_blocking (int fd, int should_block)
{
struct termios tty;
memset (&tty, 0, sizeof tty);
if (tcgetattr (fd, &tty) != 0)
{
error ("error %d getting term settings set_blocking", errno);
return;
}
tty.c_cc[VMIN] = should_block ? 1 : 0;
tty.c_cc[VTIME] = should_block ? 5 : 0; // 0.5 seconds read timeout
if (tcsetattr (fd, TCSANOW, &tty) != 0)
error ("error setting term %sblocking", should_block ? "" : "no");
}