ftdi-usb-sio-devel — FTDI USB Device Driver development discussion list

Re: [Ftdi-usb-sio-devel] 62 bytes... comes from where?

[Kentaro F. <fu...@is...>]

When I replaced USB core driver to JE driver instead of UHCI, it worked well

???

Kentaro Fukuchi

Re: [Ftdi-usb-sio-devel] 62 bytes... comes from where?

[Kentaro F. <fu...@is...>]

Hello,

Some addition to my previous mail.

Our code works well with FTDI's driver for Windows.

I have read the driver code and realized the meaning of 62 bytes.
I modified driver to dump some debug messages. We had got a lot of
TTY_OVERRUN when the baud rate was 700Kbps. I can't belive that
our machine is not fast enough to receive 700Kbps data (Pentium4 1.7GHz).

Here is a test code to receive data from usbserial.

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <fcntl.h>
#include <unistd.h>
#include <termios.h>
#include <errno.h>
#include <linux/serial.h>
#include <asm/ioctls.h>
#include <sys/time.h>

#define DATA_LENGTH 769
int fd;

int device_open(char *filename)
{
	struct termios tty;

	fd = open(filename, O_RDWR);
	if(fd < 0) return -1;

	tcgetattr(fd, &tty);
	tty.c_cflag = CS8|CREAD|CLOCAL|B38400;
	tty.c_iflag = IGNBRK;
	tty.c_oflag = 0;
	tty.c_lflag = 0;
	tty.c_line = 0;
	tty.c_cc[VTIME] = 0;
	tty.c_cc[VMIN] = 1;

	tcsetattr(fd, TCSAFLUSH, &tty);

	return 0;
}

int main(int argc, char **argv)
{
	struct serial_struct serinfo;
	unsigned char buf[800];
	int ret;
	int x, y, ymax;
	int p;
	struct timeval t1, t2;
	int total_byte = 0;
	int ttt;
	int i, j;

	if(argc != 2) {
		puts("missing device file name");
		exit(1);
	}
	if(device_open(argv[1])) {
		perror("failed to open device: ");
		exit(1);
	}

	serinfo.reserved_char[0] = 0;
	if(ioctl(fd, TIOCGSERIAL, &serinfo) < 0) {
		perror("failed to get serial info");
		exit(1);
	}
	printf("Baud_base: %d, divisor: %d\n",
			serinfo.baud_base, serinfo.custom_divisor);
	serinfo.custom_divisor=34;
	serinfo.flags |= ASYNC_SPD_CUST;
	if(ioctl(fd, TIOCSSERIAL, &serinfo) < 0) {
		perror("failed to set serial info");
		exit(1);
	}
	while(1) {
		gettimeofday(&t1, NULL);
		for(i=0; i<100; i++) {
			buf[0] = 0xff;
			ret = write(fd, buf, 1);
			if(ret != 1) {
				puts("w?");
			}
			total_byte = 0;
			while(total_byte < DATA_LENGTH) {
				ret = read(fd, buf, DATA_LENGTH-total_byte);
				printf("ret %d\n", ret);
				if(ret < 0) {
					if(errno != EAGAIN) {
						perror("");
					}
				} else {
					total_byte += ret;
				}
			}
		}
		gettimeofday(&t2, NULL);
		ttt = (t2.tv_sec - t1.tv_sec)*1000000 + t2.tv_usec - t1.tv_usec;
		printf("bps = %f\n", (float)DATA_LENGTH * 800 * 1000000/ ttt);
	}
	return 0;
}


Any suggestion are welcome.

Kentaro Fukuchi

[Ftdi-usb-sio-devel] 62 bytes... comes from where?

[Kentaro F. <fu...@is...>]

Hello,

We have developed a device employs FT8U232AM and are testing it.
I got some problem with ftdi_sio driver and I need any help.

The device has an FT8U232AM connected to an AVR CPU. When it receives
one byte, then AVR sends 770 bytes data sequentially.
When the baud rate is under 353Kbps (means custom_divisor <= 68),
there is no problem. When the baud rate is up to 700Kbps, or custom_divisor
is 34, it fails to read data from the device file.

We observed following two phenomena.

1) read() call always returns 62 bytes data.
2) when it failed to read, 691 bytes had been read, and read() system-call
   was blocked. After that, we send more one byte, and 691 bytes data returns
   again, so, AVR, FTDI chip and the driver all do not stuck. Rest 79 bytes
   should be lost.

I know those observations are too unclear and it may not be caused by the
ftdi_sio driver. But we need any help or advice.

Our question is

1) why read() returns 62 bytes at once?
2) who drops rest 79 bytes? FTDI or the driver?

We uses the latest driver (version 1.2.1).

Thanks in advance.

Kentaro Fukuchi

Re: [Ftdi-usb-sio-devel] Support of USB to FIFO parallel device

[Bill R. <br...@pa...>]

I have only tested with the 245AM so far so I am not sure what issues 
you might have. Assuming the protocol to talk to the device is the same 
at the most you should just need to change the vendor/product id.

Please note that the linux driver is not written for performance so the 
maximum data transfer rate full-duplex is likely to be only about 1Mbps 
(bits per second) depending on your host.

Stephen Noftall wrote:

>Hello;
>
>I am using FTDI's FT245BM chip with Linux on an embedded board I am
>designing.
>
>I was wondering if anyone has used this chip before? Is there any gotchas
>with it?
>
>Thanks for any heads up.
>
>Stephen Noftall
>
>
>
>
>
>-------------------------------------------------------
>This sf.net email is sponsored by:ThinkGeek
>Welcome to geek heaven.
>http://thinkgeek.com/sf
>_______________________________________________
>Ftdi-usb-sio-devel mailing list
>Ftd...@li...
>https://lists.sourceforge.net/lists/listinfo/ftdi-usb-sio-devel
>
>
>  
>

[Ftdi-usb-sio-devel] subscribe

[Tom O. <ote...@ho...>]

I want to subscibe to the ftdi-usb-sio-devel list

Tom

_________________________________________________________________
Join the world’s largest e-mail service with MSN Hotmail. 
http://www.hotmail.com

[Ftdi-usb-sio-devel] Support of USB to FIFO parallel device

[Stephen N. <ste...@lc...>]

Hello;

I am using FTDI's FT245BM chip with Linux on an embedded board I am
designing.

I was wondering if anyone has used this chip before? Is there any gotchas
with it?

Thanks for any heads up.

Stephen Noftall

[Ftdi-usb-sio-devel] Welcome/test

[Bill R. <br...@sg...>]

This should be the first message in the mailing list.

Please use this list for experiences/patches/questions about the ftdi 
usb driver for linux


-- 

Bill Ryder