Experimental Physics and
Industrial Control System

Jennings Guy <[email protected]> · Thu, 18 Apr 2013 15:55:07 -0500

Oh, and this article should help with setting up the serial number -> device udev rules

http://hintshop.ludvig.co.nz/show/persistent-names-usb-serial-devices/

On Apr 18, 2013, at 3:46 PM, Jennings Guy <[email protected]> wrote:

> I see that you are using kernel 2.6.35.10…
> 
> Kernel 3.8 has the apropriate constants defined in the ftdi driver (at least for the 0x3000 device)
> 
> http://lxr.free-electrons.com/ident?i=NEWPORT_VID
> 
> and
> 
> http://lxr.free-electrons.com/ident?i=NEWPORT_AGILIS_PID
> 
> so your 0x3000 device should work out of the box there.  Additionally, you can
> pass ONE additional vendor:product id pair to the ftdi_sio module when it loads
> so you could recognise both devices.  This should enable the ftdi_sio module to
> recognise both devices at the same time.  Note that the thing you load with modprobe
> is a 'module', not a driver, and you can only have one instance of it loaded at a time.
> 
> The longer term solution would be to get the second device ID into the linux kernel
> sources. This article
> 
> http://www.sealevel.com/support/article/AA-00524/0/How-to-configure-USB-serial-adapters-in-Linux.html
> 
> (written for a different manufacturers product) gives a reasonable level of detail for
> what needs to be done.
> 
> You can then use udev rules to match serial numbers to device nodes.
> 
> Hope this is helpful
> 
> Guy Jennings
> 
> On Apr 18, 2013, at 3:15 PM, Mark Rivers <[email protected]> wrote:
> 
>>> With no /dev/ttyUSB* device files, what happens if the second device is
>>> plugged in first?  
>> 
>> Without unplugging the devices I can unload the driver
>> 
>> Colorado:motor/iocBoot/iocWithAsyn>sudo modprobe -r -v ftdi_sio
>> rmmod /lib/modules/2.6.35.10-74.fc14.x86_64/kernel/drivers/usb/serial/ftdi_sio.ko
>> rmmod /lib/modules/2.6.35.10-74.fc14.x86_64/kernel/drivers/usb/serial/usbserial.ko
>> 
>> At this point there are no /dev/ttyUSB devices:
>> Colorado:motor/iocBoot/iocWithAsyn>ls -l /dev/ttyUSB*
>> ls: No match.
>> 
>> Now I load the driver for model 3006
>> Colorado:motor/iocBoot/iocWithAsyn>sudo modprobe -v ftdi_sio vendor=0x104d product=0x3006 debug
>> insmod /lib/modules/2.6.35.10-74.fc14.x86_64/kernel/drivers/usb/serial/usbserial.ko 
>> insmod /lib/modules/2.6.35.10-74.fc14.x86_64/kernel/drivers/usb/serial/ftdi_sio.ko vendor=0x104d product=0x3006 debug
>> 
>> It works, and ttyUSB0 exists:
>> Colorado:motor/iocBoot/iocWithAsyn>ls -l /dev/ttyUSB0
>> crw-rw---- 1 root dialout 188, 0 Apr 18 15:09 /dev/ttyUSB0
>> 
>> Now I unload the driver again:
>> Colorado:motor/iocBoot/iocWithAsyn>sudo modprobe -r -v ftdi_sio
>> rmmod /lib/modules/2.6.35.10-74.fc14.x86_64/kernel/drivers/usb/serial/ftdi_sio.ko
>> rmmod /lib/modules/2.6.35.10-74.fc14.x86_64/kernel/drivers/usb/serial/usbserial.ko
>> 
>> No /dev/ttyUSB devices:
>> Colorado:motor/iocBoot/iocWithAsyn>ls -l /dev/ttyUSB*
>> ls: No match.
>> 
>> Now I load the driver but for model 3000 this time:
>> Colorado:motor/iocBoot/iocWithAsyn>sudo modprobe -v ftdi_sio vendor=0x104d product=0x3000 debug
>> insmod /lib/modules/2.6.35.10-74.fc14.x86_64/kernel/drivers/usb/serial/usbserial.ko 
>> insmod /lib/modules/2.6.35.10-74.fc14.x86_64/kernel/drivers/usb/serial/ftdi_sio.ko vendor=0x104d product=0x3000 debug
>> Colorado:motor/iocBoot/iocWithAsyn>ls -l /dev/ttyUSB*
>> crw-rw---- 1 root dialout 188, 0 Apr 18 15:09 /dev/ttyUSB0
>> 
>> So it works.  But I don't know how to also have it recognize the second device.  As Andrew said, loading the driver again does nothing:
>> Colorado:motor/iocBoot/iocWithAsyn>sudo modprobe -v ftdi_sio vendor=0x104d product=0x3006 debug
>> Colorado:motor/iocBoot/iocWithAsyn>ls -l /dev/ttyUSB*
>> crw-rw---- 1 root dialout 188, 0 Apr 18 15:09 /dev/ttyUSB0
>> 
>> I'll look into udev.
>> 
>> Thanks,
>> Mark
>> 
>> 
>> ________________________________________
>> From: J. Lewis Muir [[email protected]]
>> Sent: Thursday, April 18, 2013 2:17 PM
>> To: Mark Rivers
>> Cc: [email protected]
>> Subject: Re: Linux USB serial questions
>> 
>> On 4/18/13 1:21 PM, Mark Rivers wrote:
>>> Thanks for all the replies I've received so far.  I am still reading and understanding them.
>>> 
>>> But now I have a more fundamental question, there is something I am not understanding about modprobe I think.
>>> 
>>> I now have 2 different USB ftdi_sio devices.  Both are manufactured by Newport but they are different models.
>>> 
>>> This is what lsusb shows for these devices:
>>> 
>>> Colorado:motor/iocBoot/iocWithAsyn>/usr/sbin/lsusb
>>> Bus 004 Device 007: ID 104d:3006 Newport Corporation
>>> Bus 002 Device 004: ID 104d:3000 Newport Corporation
>>> 
>>> Initially I plugged in the device with model=3006.  I then did modprobe to load the ftdi_sio driver for that vendor and model, and it created /dev/ttyUSB0.  It works fine, I can control it with no problems.
>>> 
>>> Now I plugged in the second device with model=3000, leaving the first device also plugged in.  The new device shows up with lsusb, as seen above.  However, it did not create a /dev/ttyUSBn for that device.
>>> 
>>> I then tried to run modprobe for the new device:
>>> 
>>>> sudo modprobe ftdi_sio vendor=0x104d product=0x3000
>>> 
>>> There is no error message.  However, it did not create a new /dev/ttyUSBn device:
>>> 
>>>> ls -l /dev/ttyUSB*
>>> crw-rw-rw- 1 root dialout 188, 0 Apr 18 13:10 /dev/ttyUSB0
>>> 
>>> There is still only device ttyUSB0, not a new ttyUSB1.
>>> 
>>> What am I doing wrong?
>> 
>> Hi, Mark.
>> 
>> With no /dev/ttyUSB* device files, what happens if the second device is
>> plugged in first?  Does it get assigned /dev/ttyUSB0?
>> 
>> Another thing to try would be to add the "debug" keyword to the end of
>> the modprobe command, and see if you get anything helpful in dmesg (or a
>> log file).
>> 
>> Lewis
>> 
> 
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System