[Asterisk-Users] Re: Snom firmwares suck <--additional datapoint to consider

[Rich Adamson]

Benny Amorsen wrote:
>>>>>> "RA" == Rich Adamson <radamson at routers.com> writes:
> 
> RA> setting. Essentially, the switch port "and" the attached device
> RA> auto negotiates at the same time, and one device "sees" what it
> RA> thinks is half duplex when the other device is in the middle of
> RA> its negotiation process. In most cases, statically defining "one"
> RA> of the two is sufficient, but to be 100% accurate from a
> RA> performance perspective, both should be statically defined.
> 
> Never ever force one device to a speed/duplex setting and leave the
> other one auto. Standard-conforming devices will switch to half duplex
> when set to autonegotiate and seeing that the other side doesn't
> answer (which it won't, when set manually).

That's a total crock. There isn't any such thing as "other side doesn't 
answer" for speed & duplex negotiation. Both ends are totally and 100% 
independent of each other, and happens way before any layer two or three 
communications functions being available.

Having worked (professionally) in hundreds of accounts over the last 15 
years or so, there are several nic card drivers miscoded in such a way 
as to only work correctly when one side is statically defined and the 
miscoded drivers only function correctly in auto-negotiate mode.

Back to the OP topic... it is entirely possible the snom firmware might 
be seeing certain type(s) of packets that its drivers don't understand, 
and most of the postings relative to the original topic simply intended 
to try disabling those that might be controllable as one step to 
eliminating possibilities.

RA