Re: [Ftdi-usb-sio-devel] howto: setting up driver to ftdi usb

[Mike R. <er...@es...>]

On Thu, 14 Nov 2002, Bill Ryder wrote:

> Does it write into /var/log/messages saying a FTDI device was detected?
> /proc doesn't tell the whole story - it will take you a usb device was
> detected.

Yeah, baby, this looks really good:

Nov 14 21:23:07 localhost kernel: hub.c: USB new device connect on bus1/1,
assigned device number 2
Nov 14 21:23:07 localhost kernel: usb.c: USB new device connect, assigned
device number 2
Nov 14 21:23:07 localhost kernel: Manufacturer: FTDI
Nov 14 21:23:07 localhost kernel: Product: USB <-> Serial
Nov 14 21:23:07 localhost kernel: usbserial.c: FTDI 8U232AM converter
detected
Nov 14 21:23:07 localhost kernel: usbserial.c: FTDI 8U232AM converter now
attached to ttyUSB0 (or usb/tts/0 for devfs)

and dmesg returns:

hub.c: port 1 connection change
hub.c: port 1, portstatus 101, change 1, 12 Mb/s
hub.c: port 1, portstatus 103, change 10, 12 Mb/s
hub.c: USB new device connect on bus1/1, assigned device number 2
usb.c: USB new device connect, assigned device number 2
usb.c: kmalloc IF d18c5900, numif 1
usb.c: new device strings: Mfr=1, Product=2, SerialNumber=0
usb.c: USB device number 2 default language ID 0x409
Manufacturer: FTDI
Product: USB <-> Serial
usbserial.c: Looking at FTDI 8U232AM Vendor id=0403 Product id=6001
usbserial.c: descriptor matches
usbserial.c: found bulk in
usbserial.c: found bulk out
usbserial.c: FTDI 8U232AM converter detected
usbserial.c: get_free_serial 1
usbserial.c: get_free_serial - minor base = 0
usbserial.c: usb_serial_probe - setting up 1 port structures for this
device
usbserial.c: FTDI 8U232AM converter now attached to ttyUSB0 (or usb/tts/0
for devfs)
usb.c: serial driver claimed interface d18c5900

So I'm doing something stupid in my calls to the driver.  Here's what I
do:

	serial = open( portname, O_RDWR | O_NOCTTY | O_NONBLOCK);
  [..error check on serial ok..]
	tcgetattr( serial, &tio);
	printf("tio.c_cc[vmin] = %d, [vtime] = %d\n", tio.c_cc[VMIN],
tio.c_cc[VTIME]);
	tio.c_iflag = 0;
	tio.c_oflag = 0;
	tio.c_cflag = CREAD | CRTSCTS | CS8 | B38400 | CLOCAL;
	tio.c_lflag = 0;

/*  use time instead of data.  If no response, something wrong.
	If there is  a response, deal with as much as possible.
*/
	tio.c_cc[VMIN] = 0;
	tio.c_cc[VTIME] = 1;
	tcflush( serial, TCIOFLUSH);
	tcsetattr( serial, TCSANOW, &tio);

   [...other stuff for me ...]
	i = ioctl( serial, TIOCMGET, &arg);
	printf("error = %d  iniitial bits are %x\n", i, arg);
	outmsg[0] = 'A';
	outmsg[1] = 'B';
	outmsg[2] = 'C';
	num = write( serial, outmsg, 3);
	if(num<0) perror("write didn't work");
	j = 0;
	while( j<10)
	{
		num = read( serial, inmsg, 63);
		if (num < 0) perror("first read didn't work");
	:
	:

and here's what I get:

tio.c_cc[vmin] = 0, [vtime] = 1
error = 0  iniitial bits are 120
Initializing com link to jtag interface

first read didn't work: Resource temporarily unavailable
number of bytes from 0x81 command -1
number of bytes from 0x81 command 3
received A@A
init_stream returns 4

>  I am making an *assumption* (possibly unwarranted) that it won't work
> because I am not aware of anyone backporting the current drivers - and
> the early versions had plenty of bugs and they only supported the
> original SIO.

Yeah, it seems like I should be able to fake my way past it tho.
I just gotta make the driver think it's doing something useful,
even tho there are no serial lines out there.

> ftdi_sio_write
>
> The trick with the devices is the original device (the SIO)  required a
> byte count on the front of the output packet, the newer devices (ie the
> 232AM and 232BM, 245 don't.
>
> So your choices are to backport the current driver, or compare the
> current driver to the driver in 2.2 and fix it manually. Certainly not
> impossible.

Oh I see, that extra byte is being sent to the device, and it's not
really needed.  Maybe I can fake that out too, but I guess talking to
the usb device directly rather than to the serial device might be simpler.

Thanks, I think it's getting closer.  Yesterday I didn't even get the "A"
back, just pure garbage.  If all I have to do is change the open() line
and it makes the driver happy, I may be in the clear.

Patience, persistence, truth,
Dr. mike