Am 24.09.2014 um 13:42 schrieb Julian Gold <[email protected]>:
> So I have been trying to write code to read data asynchronously from a serial
> port. I have tried it with /dev/ttyO1 (using UART1), and also with
> /dev/ttyUSB0 (using an FTDI USB cable in the USB port).
>
> Here is a snippet of code. This fails on both serial ports, but in different
> manners:
>
>
>
> void readSerial( int fd )
> {
> fd_set readfs;
> FD_ZERO( &readfs );
> FD_SET( fd, &readfs );
> while( 1 )
> {
> select( fd+1, &readfs, 0, 0, 0 );
> if (FD_ISSET( fd, &readfs ))
> {
> char buffer[32];
> int n = read( fd, buffer, 32 );
> printf( buffer ); fflush( stdout );
> }
> }
> }
>
>
> On /dev/ttyO1, the select() call blocks permanently, with no input being
> acknowledged. On /dev/ttyUSB0, the function prints EVERY OTHER character, so
> if I type "connect" into my PC terminal program, I will get either "cnet" or
> "onc" depending on what came before. Either way, it appears select() is
> broken because if I replace this with
>
> while((n= read(fd,buffer,32)) > 0)
> {
> printf(buffer); fflush(stdout);
> }
>
> Then I get "connect" as expected on /dev/ttyUSB0 (untried as yet on
> /dev/ttyO1). This seems like a fundamental problem with the Angstrom kernel?
No. With your program logic. You do not process the return value of the read()
system call which might be 1..32.
And, printf(buffer) is very dangerous. What would happen if you receive the
string "%s"?
Try this code:
http://git.goldelico.com/?p=gta04-rootfs.git;a=blob;f=config/root/femtocom.c;h=7f15df363f384786fad084c1217d05d00a6746e3;hb=HEAD
-- hns
--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups
"BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
