Experimental Physics and
Industrial Control System

Benjamin Franksen <[email protected]> · Mon, 11 Dec 2000 13:28:57 +0100

Andrew Johnson wrote:
> 
> Mathew Rippa wrote:
> >
> > Has anybody experienced problems with the LanGpib driver from Benjamin Franksen
> > (any version) returning the wrong ioc address? I have consulted Peter Mueller
> > (from Bessy) on this. It looks like a bug in the call to
> > "get_myaddress(&iocAddr)".
> >
> > The driver returns only 128.171.0.0 when our actual address is 128.171.90.150
> > .. As a result it incorrectly thinks our gateway is 128.171.0.59,  which is
> > actually 128.171.90.59 .  Verbose details below...
> 
> In section 2.4 of the lanGpib documentation, ie at
> http://www-csr.bessy.de/control/SoftDist/gpib/doc/lanGpib-8.html#HEADING8-0
> there is a paragraph labeled WARNING indicating that the driver does not
> use the IOC's _subnet_ mask but its _netmask_ to determine the box's IP
> address.  128.x.x.x is a Class B address, so your netmask is 255.255.0.0,
> although your local subnet mask may be 255.255.255.0 (which is irrelevent
> for this driver).  The link part of your GPIB hardware addresses need to
> have 90*256=23040 added to them.  This caught us at APS the first time
> too.

Yes. Thanks for pointing this out, Andrew. I am not so fast answering
GPIB questions these days, because my brain is so stuffed with other
things.

> There is a quick reference for IP address classes at
> http://foldoc.doc.ic.ac.uk/foldoc/foldoc.cgi?Internet+address
> 
> Ben - I suggest you revise the wording of that paragraph, and maybe add an
> example as well to make it more obvious.  Something like:
> 
> WARNING: The IP address of the gateway is determined by combining the
> value in the link field with the network part of the IOC's address, not
> the subnet part.  If the gateway's IP address is A.B.C.D and A is between
> 128 and 191, then you have a class B network, so your link address must be
> C*256+D.  If A is between 192 and 223 then you have a class C network, and
> the link address is just D.  If A is less than 128 and B is non-zero, you
> can't use this driver.

Thanks for the suggestion. I have updated the doc accordingly.

Ben

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System