cvs commit: src/sys/netgraph ng_pppoe.c

[Julian Elischer]

Gleb Smirnoff wrote:

>On Thu, Jan 26, 2006 at 11:49:08AM -0800, Julian Elischer wrote:
>J> try not to change the behaviour of the LMI and PPPoE modules too much.
>J> the modules as they were, were certified by MCI as being complient with 
>J> their
>J> PPPoE requirements.
>J> Some of the required behaviours were odd to me becasue they were not 
>J> specified by the spec.
>J> (e.g. this behaviour)
>
>Ok. Should I hide the new behavior under a sysctl, turned off by
>default? It isn't destructive however. Is there any available test
>to verify whether any check is broken with new code
>
It should be controllable, but you can enabe it by defalult..

I no longer have the specs etc. and have lost contact with the MCI guy.
(though he was a freebsd user which was cool)

>?
>
>The other change I'm planning to do is the following - if the
>original PADI had empty Service-Name, and we are servicing a
>specific Service-Name, then return remove empty one from PADO,
>returning only our specific Service-Name.
>
>J> An interesting story.. 
>J> 
>J> I went to Dallas where their test lab was, with a laptop and the test 
>J> box (actually 2 of them)
>J> (A whistle Interjet) and all the sources. When, in the first morning, 
>J> they gave me a list of
>J> tests for which they wanted changes, (i.e. it failed,) I made the 
>J> changes over the
>J> lunch hour  on my laptop, and comiled up a new kernel and asked them to
>J> retest in the afternoon. They were amazed as they had never seen 
>J> turnarund in less that 3 months
>J> (mainly from big companies).
>J> 
>J> Needless to say, it passed with Flying colours in the second try, but 
>J> according to  the paperwork
>J> it passed with flying colours on the first try (from their perspective). 
>J> Something else they
>J> had never seen.     :-)
>
>I did the draft of this change at the D-Link office, too, while their
>guy was testing different gear against current FreeBSD. :)
>
>  
>