cp210x.ko hangs on open attempt

[Roger K. Wells]

I have a RS485 serial device (Quatech SSU2-300) that when plugged
into a USB (2.0) port appears to be recognized correctly. dmesg yields:

cp210x 3-1:1.0: cp210x converter detected
usb 3-1: reset full speed USB device using uhci_hcd and address 8
usb 3-1: cp210x converter now attached to ttyUSB0

A "C" language program attempting to use this port hangs at the
open call.

This same program with the same device worked on a Centos 5.5
install on the same computer.

uname -r : 2.6.34.7-56.fc13.x86_64

The USB device works in Windows XP so I don't think there is a defect
there.

Any thoughts will be appreciated.
THX

[Rick Stevens]

On 09/29/2010 08:55 AM, Roger K. Wells wrote:
> I have a RS485 serial device (Quatech SSU2-300) that when plugged
> into a USB (2.0) port appears to be recognized correctly. dmesg yields:
>
> cp210x 3-1:1.0: cp210x converter detected
> usb 3-1: reset full speed USB device using uhci_hcd and address 8
> usb 3-1: cp210x converter now attached to ttyUSB0
>
> A "C" language program attempting to use this port hangs at the
> open call.
>
> This same program with the same device worked on a Centos 5.5
> install on the same computer.
>
> uname -r : 2.6.34.7-56.fc13.x86_64
>
> The USB device works in Windows XP so I don't think there is a defect
> there.
>
> Any thoughts will be appreciated.

Check the permissions on /dev/ttyUSB0 and make sure the user you're 
running the C program as has permissions.  You might also have a
look to see if there's an SELinux denial associated with the attempt.
It's not uncommon for udev and/or packagekit and/or SELinux to set up
incorrect permissions or not have a context rule--especially on a
device few people have around to test with (and I doubt a lot of us
have a USB<-->RS485 adapter).  :-)

If you have the source to the program you could change the open call to
a non-blocking open and print out the resulting error code:

#include <string.h>
    #include <errno.h>

extern int errno;
    if ((fdc = open("/dev/ttyUSB0", O_RW | O_NONBLOCK)) < 0)
    	printf("Can't open ttyUSB0: %s\n", strerror(errno));

You get the idea.
----------------------------------------------------------------------
- Rick Stevens, Systems Engineer, C2 Hosting          [e-mail] -
- AIM/Skype: therps2        ICQ: 22643734            Yahoo: origrps2 -
-                                                                    -
-      Cuteness can be overcome through sufficient bastardry         -
-                                         --Mark 'Kamikaze' Hughes   -
----------------------------------------------------------------------

[Alan Cox]

On Wed, 29 Sep 2010 11:55:21 -0400
"Roger K. Wells"  wrote:
> I have a RS485 serial device (Quatech SSU2-300) that when plugged
> into a USB (2.0) port appears to be recognized correctly. dmesg yields:
>
> cp210x 3-1:1.0: cp210x converter detected
> usb 3-1: reset full speed USB device using uhci_hcd and address 8
> usb 3-1: cp210x converter now attached to ttyUSB0
>
> A "C" language program attempting to use this port hangs at the
> open call.
>
> This same program with the same device worked on a Centos 5.5
> install on the same computer.
>
> uname -r : 2.6.34.7-56.fc13.x86_64
>
> The USB device works in Windows XP so I don't think there is a defect
> there.

Modern kernels properly handle carrier detect and wait so open it with
O_NDELAY if you don't want to wait for the carrier line to be asserted
and/or set the CLOCAL termios bits if you are not using the carrier and
flow control.

[Wolfgang S. Rupprecht]

Alan Cox  writes:
> On Wed, 29 Sep 2010 11:55:21 -0400
> "Roger K. Wells"  wrote:
> > I have a RS485 serial device (Quatech SSU2-300) that when plugged
> > into a USB (2.0) port appears to be recognized correctly. dmesg yields:
> >
> > cp210x 3-1:1.0: cp210x converter detected
> > usb 3-1: reset full speed USB device using uhci_hcd and address 8
> > usb 3-1: cp210x converter now attached to ttyUSB0
> >
> > A "C" language program attempting to use this port hangs at the
> > open call.
> >
> > This same program with the same device worked on a Centos 5.5
> > install on the same computer.
> >
> > uname -r : 2.6.34.7-56.fc13.x86_64
> >
> > The USB device works in Windows XP so I don't think there is a defect
> > there.
>
> Modern kernels properly handle carrier detect and wait so open it with
> O_NDELAY if you don't want to wait for the carrier line to be asserted
> and/or set the CLOCAL termios bits if you are not using the carrier and
> flow control.

One of my pet peeves is people that wire up 3-wire rs-232 (ground, tx,
rx) and then wonder why they can't open the port.  Carrier detect is a
very useful line for detecting if the rs-232 plug fell out of the
computer.  People should get in the habit of at least looping RTS->CTS,
DTR->{DSR,DCD}.

-wolfgang